Abstract

Background. Heparin-induced thrombocytopenia (HIT) is a clinicopathologic diagnosis requiring thrombocytopenia, a thrombotic or other heparin-related toxicity and demonstration of heparin-PF4 antibodies. A recently described clinical scoring system to determine the pre-test probability of HIT, the 4 T’s, appears promising in evaluation of suspected HIT when based on serotonin release assay (ASH 2003, #1963). We sought to evaluate this scoring system in conjunction with the heparin-PF4 ELISA, which is more widely available. Furthermore, we sought to determine if optical density (OD), as a correlate of heparin-PF4 antibody levels, may have clinical significance.

Methods. A retrospective review of all heparin-PF4 antibody tests (GTI, Waukesha, WI) at our institution over 3 years was performed. The clinical probability of HIT was calculated using the 4 T’s model through an extensive review of computerized medical records. The clinical probability was compared to the OD of the ELISA for heparin-PF4.

Results. There were 134 heparin-PF4 tests performed, sufficient clinical data was available for 104 patients. The distribution of low probability (LP), moderate probability (MP) and high probability (HP) was 41%, 43% and 15%, respectively. Overall, 14% received LMWH, 82% received unfractionated heparin, and 5% received no documented heparin. The indications for anticoagulation were prophylaxis in 49% and treatment in 46%. In-hospital death occurred in 25% of patients. Hematology consultation was obtained in 54% of cases. In patients who received a hematology consultation and had a high clinical probability of HIT, 78% were started on an alternate anticoagulant (p=0.002). In patients who did not receive a hematology consultation and who had a high clinical probability, only 29% received alternate anticoagulation (p=0.099). The OD of the ELISA for heparin-PF4 antibodies was 0.146 (0.066–0.552), 0.239 (0.087–2.883) and 0.396 (0.135–3.689), for LP, MP and HP groups, respectively. This was statistically significant between LP and MP and between LP and HP, (p <0.0005), but not statistically significant between MP and HP (p=0.060).

Conclusions. There is a trend toward higher heparin-PF4 antibodies as measured by OD as clinical probability of HIT increases; larger prospective studies are needed to confirm these findings. The use of alternate anticoagulation more closely reflects clinical scores when hematology consultation is obtained.

Likelihood of alternate anticoagulation based on pretest probability (4T’s) without hematology consultation

 (−) alt. anticoagulation (+) alt. anticoagulation Total 
Low 20 (95%) 1 (5%) 21 (100%) 
Moderate 16 (76%) 5 (24%) 21 (100%) 
High 5 (71%) 2 (29%) 7 (100%) 
 (−) alt. anticoagulation (+) alt. anticoagulation Total 
Low 20 (95%) 1 (5%) 21 (100%) 
Moderate 16 (76%) 5 (24%) 21 (100%) 
High 5 (71%) 2 (29%) 7 (100%) 

Likelihood of alternate anticoagulation based on pretest probability (4T’s) with hematology consultation

 (−) alt. anticoagulation (+) alt. anticoagulation Total 
Low 18 (82%) 4 (18%) 22 (100%) 
Moderate 12 (50%) 12 (50%) 24 (100%) 
 2 (22%) 7 (78%) 9 (100%) 
 (−) alt. anticoagulation (+) alt. anticoagulation Total 
Low 18 (82%) 4 (18%) 22 (100%) 
Moderate 12 (50%) 12 (50%) 24 (100%) 
 2 (22%) 7 (78%) 9 (100%) 

Author notes

Corresponding author