Background: Corrected count increment (CCI) and percent platelet recovery (PPR) are measures of response to platelet transfusion that "correct" the count increment for blood volume and number of platelets transfused. Their potential for data distortion is described, and a regression analysis is suggested that is more informative and avoids the inherent problems associated with using ratios as outcome measures.
Study design and methods: Data from the first platelet transfusion for 585 patients from the Trial to Reduce Alloimmunization to Platelets (TRAP) were used to model methods of analyzing posttransfusion platelet response.
Results: By linear regression analysis, unfiltered platelet components gave a greater posttransfusion increment on average (p = 0.001), but filtered platelets gave a greater increment per platelet transfused (p = 0.003). In contrast, CCI and PPR showed no difference between filtered and unfiltered platelets (p = 0.36 and p = 0.29, respectively) because they combined the effects of dose, filtration, and patient size. Slightly fewer patients are required for a study analyzed by regression analysis.
Conclusion: Regression analysis of posttransfusion platelet increments should be used instead of CCI or PPR to compare the efficacy of platelet components. CCI and PPR should not be used to define platelet refractoriness as a study outcome, because these measures are biased in favor of platelet preparation techniques that provide fewer platelets.