Striving to HIT the Bullseye When Diagnosing Heparin-Induced Thrombocytopenia

Background: The 4T score is a validated tool to assess the pretest probability of heparin-induced thrombocytopenia (HIT), with the negative predictive value of a low 4T score (≤ 3) reported consistently above 90%. We performed a retrospective data review of patients aged ≥18 years who had heparin/PF4 antibody testing ordered between January and August 2016 at an academic medical center. 4T scores were calculated retrospectively from data in the electronic medical record (EMR). We then prospectively instituted a mandatory data-driven algorithm utilizing an electronic 4T calculator within the EMR prior to heparin/PF4 antibody testing.

Aims: To reevaluate the utilization of laboratory testing for HIT after implementation of a data driven algorithm.

Methods: In December 2016, 4T score calculation was implemented in the EMR as a requirement prior to ordering heparin-PF4 antibody testing. We then compared the frequency and results of heparin/PF4 antibody testing in adult inpatients aged ≥18 during the post-intervention period (December 2016 to June 2017) with the pre-intervention period (January to August 2016). 4T scores were calculated retrospectively by a hematologist from data in the EMR and compared to the 4T score entered by the ordering provider.

Results: During the pre-intervention period, 161 heparin/PF4 antibody tests were performed, of which 131 (81%) were negative (optical density <0.40), 27 (17%) positive and 3 (2%) equivocal. During the post-intervention period, 105 heparin/PF4 antibody tests were performed, of which 86 (82%) were negative (optical density <0.40), 15 (14%) positive and 4 (4%) equivocal. The 4T score was significantly higher in patients with positive/equivocal heparin/PF4 antibody values compared to those with negative values (p<0.0001). ROC curves demonstrated that the 4T score was highly predictive of a positive or equivocal heparin/PF4 antibody value (AUC 0.705, 95% CI: 0.553, 0.857). Intensive care units ordered the majority (65%) of heparin-PF4 antibodies both pre- and post-intervention. The percentage of heparin-PF4 antibodies ordered with a low 4T score (≤3) was 67% pre-intervention and 57% post-intervention, a difference which trended towards, but did not reach statistical significance (p=0.130 by Chi-square test). 4T score calculation was discordant between the ordering provider and hematologist in 67% of patients.

Conclusions: Mandating 4T score calculation prior to heparin/PF4 antibody testing did not significantly decrease the number of tests ordered in the setting of a low 4T score within the first 6 months post-intervention. We plan to review additional data over the next several months and will continue education on accurate 4T score calculation and testing for HIT at our academic medical center.