Objectives: To study the variation in platelet count in critically ill children and correlate its association with their outcomes. Methods: This was a prospective cross-sectional analysis of 150 critically ill children admitted in PICU of our tertiary care hospital over period of 1 year. Laboratory data was collected with daily platelet counts from day of admission till death or discharge whichever was earlier in all patients. The study population was grouped as thrombocytopenic (platelet count < 150 x 10 9/L) and nonthrombocytopenic. They were compared with each other with respect to laboratory parameters and risk factors. Survivor and non survivors were compared with variation in platelet count. Decline in platelet count was correlated with mortality. Chi-square test, median test, ROC curve and forward stepwise binary logistic regression was used for statistical analysis. Results: Forty eight (32%) children had thrombocytopenia. They had significantly higher mortality [14(29.16%) vs 14(13.72%)] and bleeding tendency [13(28.08%) vs 4(3.92%) than non- thrombocytopenic children. Admission thrombocytopenia was not found to be risk factor for mortality. Though survivors and non- urvivors had decline in platelet count in first four days, non-survivors had significantly higher drop. Platelet counts decline >30% at 72 hours was independent risk factor (odds ratio 4.126) for mortality with high sensitivity(91.7%) and specificity (84.4%) with area under ROC curve of 0.898 which was associated with PRISMII. Conclusion: Decline in platelet count can be used as prognostic marker of poor outcome in critically ill children. 

Keywords: Thrombocytopenia; Decline in Platelet Count; Paediatric Intensive Care Unit.