37_C incubation phase.

In the case presented in Figure 129, all reactions occurred in the AHG phase with strengths of both 2_ and 3_. Multiple antibodies or a single antibody showing dosage should be considered. Does the serum reactivity match any of the remaining specificities? When a single alloantibody is present, the pattern of reactivity usually matches a pattern exactly. In our example, the serum reactivity perfectly matches the Fya pattern. The serum gave uniform positive results with all Fyapositive cells (1, 3, 4, 6, 7, 10, and 11) and negative results with all of the Fya-negative cells (2, 5, 8, and 9). The reason for the variation in strength in the AHG phase is due to dosage; all cells yielding 2_ reactions were heterozygous for Fya, and all cells yielding 3_ reactions were homozygous for Fya. Are all commonly encountered RBC antibodies ruled out? As mentioned above, in this case anti-E, -Kpa, and -Jsa were not ruled out. None of the cells on this panel were positive for Kpa or Jsa, so they cannot be the cause of the observed reactions. These two antigens are characterized as lowfrequency antigens occurring in less than 5 percent of the population. Antibodies to low-frequency antigens are uncommon because of the small chance of being exposed to the antigen; therefore, it may not be necessary to rule them out. In contrast, if a commonly encountered antibody is not ruled out, it is important to test selected cells that will rule out the presence of the antibody. In this example, anti-E still needs to be ruled out. A negative result when testing an E_ e_ FyaRBC sample would exclude anti-E (see section on selected cell panels below). Is the autologous control (last row in panel antigen profile) positive or negative? In the example shown, the autocontrol is negative, indicating that the positive reactions are caused by alloantibody, not by autoantibody. The presence of autoantibodies complicates the process of antibody identification and is discussed briefly in Special Problems in Antibody Identification. Is there sufficient evidence to prove the suspected antibody? Conclusive antibody identification requires testing the patients serum with enough antigen-positive and antigennegative RBC samples to ensure that the pattern of reactivity is not the result of chance alone. Testing the patients serum with at least three antigen-positive and three antigennegative cells (also known as the 3 and 3 rule) will result in a probability (P) value of 0.05.18 A P value is a statistical measure of the probability that a certain set of events will happen by random chance. A P value of 0.05 or less is required for identification results to be considered valid, and it means that there is a 5 percent (1 in 20) chance that the observed pattern occurred for reasons other than a specific antibody reacting with its corresponding antigen. Stated another way, it means that the interpretation of the data will be correct 95 percent of the time. When multiple antibodies are present, the 3 and 3 rule must be applied to each specificity. For example, if anti-K and anti-E are both suspected, the 3 and 3 rule would be fulfilled if three K_E_ cells reacted negatively, three K_E_ cells reacted positively, and three K_E_ cells reacted positively. Other researchers have derived formulas where P 0.05 is fulfilled with two positive cells and three negative

cells19 or two positive cells and two negative cells.20 Testing of RBCs selected from other panels is necessary when inadequate numbers of antigen-positive or antigen-negative cells are tested. In our example, the patients serum reacted with seven Fya-positive cells (1, 3, 4, 6, 7, 10, and 11) but did not react with four Fya-negative cells (2, 5, 8, and 9). As a result, selected cells are not required in this case, and the identification of anti-Fya is conclusive. Is the patient lacking the antigen corresponding to the antibody? Individuals cannot make alloantibodies to antigens that they possess; therefore, the last step in identification studies is to test the patients RBCs for the corresponding antigen. A negative result is expected and indicates that identification results are correct. If the patients RBCs are positive for the corresponding antigen, misid