We often "get away" with transfusing unmatched RBC because the incidence of unexpected antibodies in random patients experiencing medical emergencies is unknown but relatively low, e.g., ~5% is sometimes cited but with little solid evidence.
Incidence may vary according to several factors, e.g., genetic disposition, patient population's underlying diseases, number of prior transfusions, gender (females may get exposed to foreign antigens via feta-maternal bleeds as well as transfusion), concordance of antigen phenotypes of patients vs blood donors in a given locale.
In general, the incidence increases with number of transfusions but to a limit (i.e., most antibody producers respond within the first 3-4 transfusions); and the incidence in transfusion-dependent patients such as those with thalassemia is very high.
Regardless of likelihood, transfusing unmatched blood to a patient with unexpected antibodies can result in a serious hemolytic transfusion reaction.
However, in some case scenarios immediate transfusion is necessary. The transfusion service staff should alert the medical director, who can discuss options with clinical staff.
The medical director will generally talk to the staff attending the patient and indicate that, if possible, they should hold off transfusion. But if it is a case of massive bleeding where exsanguinating hemorrhage is likely, it is better to give some blood and monitor for a delayed hemolytic transfusion reaction than to let the patient bleed to death.
Transfusing when bleeding is brisk will result in much of the autologous and incompatible blood bleeding out, with the possibility of a delayed hemolytic reaction once the patient's antibody rebounds and destroys still present antigen-positive donor red cells. For example, see:
Dutton RP, Shih D, Edelman BB, Hess J, Scalea TM. Safety of uncrossmatched type-O red cells for resuscitation from hemorrhagic shock. J Trauma. 2005 Dec;59(6):1445-9.
Some transfusion services also try to minimize the risk of unmatched blood by typing their emergency supply of O Rh negative RBCs for the K antigen, since anti-K is a relatively common clinically significant antibody.
Bottom line:Unmatched blood should only be given when the patient's life is at risk if transfusion is delayed. Such events are rare. In true emergencies, unmatched blood may be warranted after consultation with the TS medical director.