DIAGNOSTIC ACCURACY OF 4TS SCORE COMPARED TO HEPARIN-INDUCED THROMBOCYTOPENIA EXPERT PROBABILITY SCORE FOR IDENTIFYING HEPARIN-INDUCED THROMBOCYTOPENIA IN A MEDICAL ICU

Abstract

SESSION TITLE: Tuesday Abstract Posters SESSION TYPE: Original Investigation Posters PRESENTED ON: 10/22/2019 01:00 PM - 02:00 PM PURPOSE: Heparin-induced thrombocytopenia (HIT) is challenging to diagnose in critically ill patients due to a higher incidence of acute thrombocytopenia. Underdiagnosis of HIT is potentially fatal, but overdiagnosis leads to unnecessary exposure to non-heparin anticoagulation. The 4Ts score was developed to aid in diagnosis but has limitations in the intensive care unit (ICU). The HIT expert probability (HEP) score is an alternative scoring system shown to have increased diagnostic accuracy in ICU patients in a subgroup analysis. We aim to validate these findings in a purely medical ICU. METHODS: 4Ts and HEP scores were retrospectively calculated for 91 patients with suspected HIT seen in an academic, tertiary care center medical ICU between January 2017 and April 2018. All patients had serotonin release assays (SRA) ordered. SRA positivity was used as the gold standard for HIT diagnosis. Scorers were blinded to both SRA and calculated results of the alternate scoring system. 4Ts scores equal to or greater than four and HEP scores equal to or greater than three represented positive scores. Sensitivity, specificity and likelihood ratios were calculated. McNemar’s test was used to evaluate differences in score results, sensitives and specificities. For significant differences, 95% confidence intervals were calculated using the Wald method. All statistical analyses used a significance level of 0.05. RESULTS: Three of our 91 patients had positive SRAs. The 4Ts score had a sensitivity of 0.67 (0.09-0.99), specificity of 0.45 (0.35-0.56) and a positive likelihood ratio (LR+) of 1.22 (0.54-2.78). The HEP score had a sensitivity of 1.00 (0.29-1.00), specificity of 0.69 (0.59-0.79) and a LR+ of 3.26 (2.83-4.46). Receiver operative characteristic (ROC) curve could not be created due to the low number of positive HIT cases. CONCLUSIONS: The accuracy measures of the HEP score were superior to the 4Ts score among medical ICU patients. The study was limited by single center status, small sample size of true positive HIT, retrospective design and having less experienced physicians conducting scoring, which has been suggested to favor HEP scoring in a previous study. Future studies continuing to look at this population are needed. CLINICAL IMPLICATIONS: The HEP score may have better diagnostic accuracy for HIT than the 4Ts score in the medical ICU population. DISCLOSURES: No relevant relationships by Thomas Iden, source=Web Response No relevant relationships by Audrey Jernigan, source=Web Response No relevant relationships by Markos Kashiouris, source=Web Response No relevant relationships by Victoria Okhomina, source=Web Response No relevant relationships by Xian Qiao, source=Web Response No relevant relationships by Kathleen Waybill, source=Web Response