The Declining Risk of ABO Incompatibilities: Twelve Years of Hemovigilance in France.

Background: ABO incompatibilities (ABOi) usually result from the failure to comply with Standard Operating Procedures (SOP). Continuous training (CT) is the main way to ensure their proper use in ultra-high security systems like civil aviation but little is known for transfusion. In France, SOP include a Beth-Vincent test at the bedside before the transfusion of Packed Red Cells (PRC). At the hospitals, Hemovigilance Officers (HO) are in charge of transfusion safety along with the settlement of traceability and the notification of all Adverse Effects (AE). If SOP and CT are requested by Law since 1993 for blood banks, they are only advised for hospitals even though the bedside Beth-Vincent test requires a special training. However, HO have more and more transposed the model of blood banks in hospitals especially since 2000. Objective: to describe time-trends of ABOi observed in e-fit, the French Hemovigilance database. Population and methods: All AE with “acute ABOi” as the final diagnostic were considered. Time-trends were studied according to the Box and Jenkins model (ARMA).

Results: From 07/01/1994 to 12/31/2005, 79 106 AE were registered in e-fit, corresponding to about 30,000,000 Labile Blood Products (LBP) transfused. 304 (0.4%) were ABOi (1/107,850 LBP transfused). There were 183 men and 121 women with 81.3% of whom more than 40 years old. 205 patients (67.4%) received only PRC, 73 (24.0%) Platelet Concentrates (PC), 22 (7.2%) Fresh Frozen Plasma (FFP) and 4 (1.3%) more than a single LBP. 18 ABOi were fatal, possibly (4 cases) probably (4 cases) or definitely (9 cases) associated with a PRC, and 1 doubtfully with a FFP. Time-trends displayed 3 periods: (1) from 1994 to 1997 ABOi increased concurrently with the frequency of all AE notifications, (2) from 1997 to 2000 they reached a steady-state with 35 ABOi/year on average (1/70,000 LBP), (3) since 2001 they regularly decreased of 14%/year on average (13 ABOi for 2005 = 1/198,450 LBP). The localization of the failure to comply with SOP (i.e. hospital, blood bank or both) was studied for PRC. From 1997 to 2005, the liability of blood banks alone or along with hospitals remained stable (resp 0.42 and 2.58/year on average) whereas those of hospitals dramatically decreased (16.25/year from 1997 to 2000 and 8.6/year from 2001 to 2005).

Conclusion: (1) ABOi are now 3 times less frequent than in 2000 ; (2) a notification bias is unlikely because both hospitals and blood banks are aware of traceability data that include the ABO phenotype for both the recipient and the LBP issued ; (3) CT applied on a nationwide basis seems to be the main factor to reduce ABOi in hospitals ; (4) a residual risk related to human errors seems to exist ; (5) the way to overcome that risk remains a matter of debate.