RESUMO
PURPOSE: Results of a study to determine the utility of combining laboratory values and clinical probability scores to improve the detection of heparin-induced thrombocytopenia (HIT) are reported. METHODS: In a retrospective, single-site, chart review-based investigation, 156 cases in which patients with suspected HIT had positive results on a widely used diagnostic test (the anti-heparin/platelet factor 4, or anti-PF4, assay) were identified; in all cases, the blood specimens had been sent to a reference laboratory for confirmation of HIT via serotonin release assay (SRA). After investigator scoring of the clinical probability of HIT in each case by the 4T's method, a multiple logistic regression model was used to evaluate the combined effect of 4T's scores and anti-PF4 assay values in predicting SRA results. RESULTS: 4T's scores indicating an intermediate or high probability of HIT combined with high anti-PF4 test values (i.e., optical density [OD] value of ≥1.4) were strongly predictive of a positive SRA result, as were high-probability 4T's scores alone. Low-probability 4T's scores alone or in combination with anti-PF4 OD values of <1.4 were highly correlated with negative SRA results. Controlling for potential confounding factors, logistic regression analysis indicated that the 4T's score was a better predictor of SRA results than the anti-PF4 test value. CONCLUSION: The combination of anti-PF4 OD values and 4T's scores accurately predicted SRA results, suggesting that the SRA may not be necessary to confirm HIT in patients with a relatively low 4T's score and a low anti-PF4 OD value.