OA3-AM24-MN-08 - Plasma Transfusion Practice: A Multi-Center Audit of 10,008 Admissions with Plasma Transfusionat 23 Hospitals

Guidelines recommend that plasma transfusions be limited to patients with an abnormal coagulation profile with either active bleeding or a planned invasive procedure. A multi-center retrospective audit showed that 78% of plasma transfusions were either administered outside of recognized indications or under-dosed (1-2 units in an adult). A large portion of these transfusions were given to correct a mildly prolonged international normalized ratio (INR), despite evidence that plasma does not correct mildly prolonged INRs. Thus, patients were exposed to potential harms of plasma, without benefit. The objective of this study was to characterize plasma transfusion practices across 23 hospitals in Canada using a comprehensive health database and to determine the variability of plasma utilization and pre-transfusion INRs across centers.

Study

Design/Methods:

We performed a multi-site, retrospective observational cohort study. Adult inpatients (age ≥ 18) at hospitals in the GEMINI hospital data sharing network between January 1, 2017, and June 30, 2022, were included. GEMINI is a health database that includes laboratory and clinical data from academic and community hospitals.  Hospital sites with missing plasma transfusions or INRs were excluded (9 sites). Patients undergoing plasma exchange were excluded. Data collected included patient demographics, laboratory tests, plasma transfusion information (dose, issue time and location), and in-hospital events.

Results/Findings:

We identified 861,212 hospital admissions during the study period, with 10,008 eligible hospital admissions for patients who received plasma transfusions across 23 hospitals.  The mean age of plasma recipients was 61.17 (SD 16.46), and 63.4% were male. Compared to the 851,204 admissions for patients without a plasma transfusion, admissions with a plasma transfusion (n=10,008) had a higher rate of in-hospital mortality (33.7% versus 7.0%) and intensive care unit admission (87.0% versus 23.6%) and a longer hospital length of stay (median 13.01 [IQR 6.51, 26.76] versus median 4.94 [IQR 2.38, 9.81]). There was significant heterogeneity across the 23 hospitals in the pre-transfusion INR; 17.40% of hospitals had a mean pre-transfusion INR of < 1.8, representing 17.12% of the plasma transfusion events (Figure 1). The mean number of plasma units transfused per 1000 inpatient days across all sites was 210.45 (SD 119.20), with a range of 46 to 592 units.  

Conclusions:

This retrospective study demonstrated significant variability in plasma transfusion practices across 23 Canadian hospitals, with substantial heterogeneity in the number of plasma units transfused per 1000 inpatient days and a large range of mean pre-transfusion INRs across sites. Further knowledge translation efforts are required to encourage appropriate plasma utilization.