PS1524 WHEN IS A CT NEEDED TO DIAGNOSE ACUTE PULMONARY EMBOLISM IN INDIVIDUALS WITH SICKLE CELL DISEASE

Abstract

Background:Sickle cell disease (SCD) is a hypercoagulable state that results in an increased incidence of thromboembolic events including deep venous thrombosis and pulmonary embolism (PE). However, the diagnosis of acute pulmonary embolism (APE) in individuals with SCD is more difficult due to other potential pulmonary complications such as acute chest syndrome, vaso‐occlusive pain, or an exacerbation of pulmonary hypertension. A D‐dimer is also unreliable in SCD as it is often elevated at baseline. Validated diagnostic tools, such as the Wells criteria are used to determine the likelihood of a healthy person (without SCD) having an APE but these have not been tested in individuals with SCD. Instead, adults with SCD who present to the hospital with chest pain often undergo lung imaging as the first line test although many other pathologies could mimic the presenting symptoms of an APE. Thus, there is often unnecessary exposure to radiation and intravenous contrast for individuals with SCD presenting with chest symptoms in the acute setting.Aims:The primary aim of this study was to evaluate the positive predictive value of the Wells criteria in predicting APE in SCD using a retrospective cohort to potentially identify a new, more specific diagnostic algorithm.Methods:This is a retrospective case cohort study. Index cases were individuals with SCD (confirmed) who underwent a CT of the chest with IV contrast (to assess for a pulmonary embolism) in the emergency department, outpatient clinic or inpatient setting from January 1, 2013 to October 1, 2017. Each encounter was evaluated for demographic data, presenting symptoms, laboratory assessments, final diagnosis and scored using the Wells criteria. Charts were excluded if CT scans were obtained for any other reason than to evaluate for APE or if the patient had did not have SCD.Results:There were 349 charts that fit the inclusion criteria. Most individuals were > 18 years (97%) and female (62%). HbSS was the most common phenotype (80%), followed by HbSC (15%) and HgbS‐Beta thalassemia (6.5%). The overall incidence of APE was 9.7%. The most common final diagnosis was vaso‐occlusive crisis (60%) and/or acute chest syndrome (18%). For patients in this cohort with an APE, the average Wells score was 5.14 (p < .001). Evaluating specifics of the Wells criteria, “signs and symptoms of a DVT” and “PE most likely diagnosis” were additionally statistically significant in determining an APE (p < .020 and p < .001). HbSS genotype (p < .030) was also statistically significant. There was no statistical difference between cases and controls for oxygen saturation, heart rate, hemoglobin, shortness of breath, age or sex. A Wells score >4 provided the best sensitivity (73.5%) and specificity (72%) in our sample. A Wells score of >/5 caused the sensitivity to decline significantly, making the three‐tier model less effective in our population.Summary/Conclusion:The Wells criteria can adequately predict who should undergo a CT with IV contrast to evaluate for APE. When using the Wells criteria, the two‐tier model allowed for the best sensitivity and specificity in our population. Logistic regression analysis did not reveal any variables that increased the sensitivity or specificity to a more specific algorithm than what was attained with the Wells Score. This study is being replicated at other institutions with large sickle cell populations to validate these findings.