Development and Validation of a Predictor Score of Persistent Thrombocytopenia in Primary Immune Thrombocytopenia in Children.

Aim: To develop a prognostic score predictive of persistent thrombocytopenia in acute primary immune thrombocytopenic (PIT) in children.

Methods: Two retrospective cohorts of PIT collected from two pediatric departments over a similar period of 25-year. One cohort served for the derivation of the score and the second served for the external validation. Inclusion criteria were: PIT age between 2 months and 14 years and persistent thrombocytopenia less than 100,000 / mm3 after 6 months of follow-up. The data was analyzed using statistical software SPAD version 5.5. The predictive index was done through a discriminant analysis.

Results: 219 cases were collected in the two studies, respectively: 149 cases (mean age = 5.7 years, sex ratio M / F = 1.48) from the derivation population and 70 cases from the validation population (mean age = 5.4 years, sex ratio M/F = 1.5). The developed score had 6 parameters: beginning onset, failure of the first line treatment, the presence of a triggering factor, age, platelet rate at 1 month and platelet rate at 3 months. The sum of the different coefficients for each patient provides a result convertible to probability of persistence of thrombocytopenia after 6 months. The ROC curve (Receiver operating characteristic) showed a good specificity (93%) and a good sensitivity of (90%) for a threshold of 35. The external validation study showed a good specificity e (100%) and good sensitivity (83.3%).

Conclusion: This prognostic clinical score validated by an external study requires further prospective studies to improve validation level and to explore its practical utility.