Pretest probability of HIT (the 4 T's)

Thrombocytopenia —

Platelet count fall >50 percent and nadir >20,000: 2 points
Platelet count fall 30 to 50 percent or nadir 10 to 19,000: 1 points
Platelet count fall <30 percent or nadir <10,000: zero points

Timing of platelet count fall —

Clear onset between days 5 and 10 or platelet count fall at ≤1 day if prior heparin exposure within the last 30 days: 2 points
Consistent with fall at 5 to 10 days but not clear (eg, missing platelet counts) or onset after day 10 or fall ≤1 day with prior heparin exposure within the last 30 to 100 days: 1 point
Platelet count fall at <4 days without recent exposure: 0 points

Thrombosis or other sequelae —

Confirmed new thrombosis, skin necrosis, or acute systemic reaction after intravenous unfractionated heparin bolus: 2 points
Progressive or recurrent thrombosis, non-necrotizing (erythematous) skin lesions, or suspected thrombosis which has not been proven: 1 point
None: zero points

Other causes for thrombocytopenia present —

None apparent: 2 points
Possible: 1 point
Definite: zero points

Test interpretation — A score is determined for each of the four above categories, resulting in a total score from zero to 8. Pretest probabilities for HIT are, as follows:


zero to 3: Low probability
4 to 5: Intermediate probability
6 to 8: High probability

Among 111 patients with a low pretest probability of HIT using this scoring system, only one had clinically significant HIT antibodies (0.9 percent). In contrast, the overall rate of clinically significant HIT antibodies was 11.4 and 34 percent in those with intermediate and high scores, respectively.

A significant concern with these findings is that there was substantial variability in the rate of clinically significant HIT antibodies at the two centers in patients with intermediate (29 versus 8 percent) or high scores (100 versus 21 percent). A number of methodologic and center-specific factors may have contributed to these differences [119].

These initial results were confirmed at two other centers, in which the incidences of a positive rapid ELISA immunoassay for HIT antibodies were 1.6 and 4 percent and that of a positive serotonin release assay or HIPA assay was zero percent for a total of 458 patients with suspected HIT and a low probability on the 4 T's test [108,120].

Accordingly, laboratory testing for HIT might reasonably be limited to patients with an intermediate or high pretest probability, since those with a low pretest probability are at very low risk of having clinically significant HIT antibodies (ie, <5 percent) [7,108,120].