Repeat Borderline or Negative Platelet Factor-4 ELISA May Yield Positive Results in Patients with Clinical Suspicion of Heparin-Induced Thrombocytopenia.

Heparin-induced thrombocytopenia (HIT) is mediated by antibodies to heparin-platelet factor 4 (PF4) complexes which lead to platelet activation and aggregation with or without thrombosis. HIT diagnosis relies on clinical and serologic grounds. Although the detection of PF4 antibodies alone is not diagnostic of HIT without thrombocytopenia (relative or absolute) and/or thrombosis, a negative result does not exclude the diagnosis. Thus, it is expected that initially negative ELISAs may turn positive on repeat testing. We reviewed all PF4 ELISA (GTI Diagnostics, Waukesha, WI) assay reports generated at our institution from January 2003 to May 2005 in search of patients who had more than one test performed in this time-period. Of 1,128 tests performed, 181 (16%) were positive, 95 (8%) were borderline, and 852 (76%) were negative. Borderline results were considered those with an OD₄₀₅ between 0.3 and 0.4 (cutoff value for positive) or two times that of the negative control of each run. No patients were evaluated using the serotonin-release assay. Nine of 70 patients (13%) with initially negative or borderline ELISAs had a positive result upon repeat testing from 2 to 63 days later (Table). All were inpatients, and 6 of 9 were from surgical specialties, mostly cardiac (5 of 6). Five of 8 patients (62.5%) with an initial positive PF4 ELISA result had either a negative or a borderline subsequent test from 7 to 225 days later (Table). The mean platelet count of the patients that had a nondiagnostic ELISA followed by a positive result was 98 x 10⁹/L (N = 9) at the time of the initial test compared with 137 x 10⁹/L at subsequent testing (p = 0.43 by paired Student’s t test). Of note, 2 patients had a normal platelet count when the ELISA result was positive. None had evidence of thrombosis. Borderline test results tended to be repeated sooner (mean of 16 days) than negative ones (mean of 21 days). We suspect that this difference reflects the clinicians’ high index of suspicion for HIT. Our findings suggest that in patients with a high pre-test probability for HIT (status post cardiac surgery with thrombocytopenia), there should be a low threshold for repeating an initial negative or borderline ELISA assay, thereby increasing detection and minimizing complications from this potentially fatal, yet common clinical entity.