Refining Th Predictive Value of Secretory Phospholipase A2 In Sickle Cell Disease Patients with Acute Chest Syndrome

Abstract 846

Elevation of the serum level of secretory phospholipase A2 (sPLA2) has previously been reported to predict impending acute chest syndrome (ACS) in sickle cell disease (SCD) patients admitted with pain. The Sickle Cell Disease Clinical Research Network initiated the PROACTIVE Feasibility Study to 1) assess the feasibility of screening, randomizing and treating with simple transfusion eligible patients and 2) assess the predictive utility of sPLA2 in screening for imminent ACS, characterize the distribution of sPLA2 values in sickle cell patients hospitalized with pain, and define a cut-off point for predicting ACS to optimize its reliability, sensitivity and specificity in predicting ACS. All SCD patients (hemoglobin SS, SC and SBeta-thalassemia) admitted for pain at a participating center were to be screened for eligibility; patients with a diagnosis of ACS at entry were not eligible. sPLA2 levels were drawn daily on consenting patients for up to three days during the admission. All patients had at least one CXR (mandated at 72 hr if not done for clinical indications) during the course of their hospital stay to determine if they had developed a new infiltrate, considered diagnostic for ACS. Patients who developed T>38°C, a sPLA2 level > 100ng/ml and had a normal CXR were eligible for randomization to observation or simple transfusion; the remaining patients were followed in an Observation arm. Of 421 patients potentially eligible for screening, 238 were enrolled in the study with 10 of these patients randomized (4 to transfusion 6 to standard care). Of these, 203 patients had from one to three sPLA2 levels drawn prior to an ACS diagnosis or did not develop ACS; if transfused, they received blood only after a diagnosis of ACS was made. Twenty-two of these patients (11%) developed ACS. Analysis of the maximum sPLA2 levels prior to ACS diagnosis revealed that a threshold of 45 ng/ml was the most accurate level to predict ACS (accuracy: 73%, sensitivity: 73%, specificity: 73%). The positive predictive value (PPV: 25%) was low due to the relative infrequency of ACS events. When sPLA2 levels were analyzed in children versus adults, sPLA2 was more accurate for adults in predicting ACS with a level of 65ng/ml having accuracy: 88%, sensitivity: 44%, specificity: 95%, PPV: 57%. The addition of a requirement for fever did not improve the accuracy of sPLA2 in either adults or children. Further analysis of sPLA2 levels with various clinical and laboratory data is ongoing to improve the PPV of sPLA2. These results indicate that sPLA2 can predict ACS with good sensitivity, but may require additional parameters to be useful in clinical management of SCD patients admitted with pain and at risk for ACS. Supported by the NHLBI.