Delayed tirofiban-induced thrombocytopenia: two case reports


Dr Annmarie Bosco, Haematology Registrar, Department of Haematology, South Eastern Area Laboratory Services, Level 4, Campus Centre, Prince of Wales Hospital, Randwick, NSW 2031, Australia.
Tel.: +61 2 93829004; fax: +61 2 93829116; e-mail:

Tirofiban-induced thrombocytopenia is a well-known although uncommon complication of therapy [1]. Delayed thrombocytopenia has been reported in association with abciximab, but not with tirofiban [2]. We describe two cases of delayed tirofiban-induced thrombocytopenia in which tirofiban-dependent platelet antibodies were identified.

Patient 1, an 80-year-old man, was admitted to a peripheral hospital with a non-ST elevation myocardial infarction and was transferred to our institution for coronary angiography. The platelet count on admission was 240 × 109 L−1. Angiography showed severe triple vessel disease. An apical left ventricular thrombus was noted on a transthoracic echocardiogram. Intravenous heparin was commenced. He had significant ongoing chest pain, requiring the addition of tirofiban. He proceeded to coronary artery bypass grafting 4 days after admission. The platelet count on day 4 after initial tirofiban exposure was 122 × 109 L−1. This fell progressively, and reached a nadir of 1 × 109 L−1 on day 9. The patient was commenced on danaparoid and intravenous immunoglobulin, and received a platelet transfusion.

Patient 2, a 67-year-old man, was admitted with a non-ST elevation myocardial infarction and commenced on intravenous heparin and tirofiban. The initial platelet count was 212 × 109 L−1. Percutaneous coronary angioplasty and stent placement to the left anterior descending artery were performed. Both tirofiban and heparin were discontinued 24 h after the procedure. Clopidogrel and aspirin were continued. On day 5 after initial tirofiban exposure, the platelet count was 199 × 109 L−1. By day 7, the platelet count was 13 × 109 L−1. Clopidogrel was withheld and the patient was commenced on corticosteroids temporarily pending investigations.

In both cases, heparin-induced thrombocytopenia was excluded by both ELISA and serotonin release assay. Tirofiban-dependent antiplatelet antibodies were demonstrable in both cases by two separate methods. Drug-dependent antibodies to platelets were demonstrated in the presence of tirofiban, using flow cytometry techniques (Fig. 1a). A monoclonal antibody immobilization of platelet antigen assay confirmed the presence of antibodies to glycoprotein (GP) IIb/IIIa, detected only with the addition of tirofiban. Both these methods have been described in detail previously [1].

Figure 1.

(a) Platelet counts of patients 1 and 2 following tirofiban administration. Counts from day 3 (patient 1) and day 6 (patient 2) after initial exposure are not available. (b) Donor platelets were incubated with patient serum and 1 mcg mL−1 of tirofiban and washed. Platelet-bound immunoglobulin was detected by flourescein-labeled antihuman IgG with positive results defined as mean platelet fluorescence intensity twice that of platelets processed in the absence of tirofiban. Positive control with known tirofiban-dependent antibodies not shown.

The platelet count of patient 1 on day 13 after tirofiban exposure was >50 × 109 L−1 and subsequently returned to normal. The platelet count of patient 2 recovered to >50 ×109 L−1 by day 11. We concluded that both these patients had a delayed tirofiban-induced thrombocytopenia, in which the nadir of the platelet counts occurred more than 6 days after first exposure to tirofiban (Fig. 1b).

These patients represent the first reported cases of delayed thrombocytopenia associated with tirofiban. GP IIb/IIIa inhibitor-induced thrombocytopenia is mediated by pre-existing autoantibodies (either naturally occurring or secondary to prior drug exposure) that bind after drug-induced conformational changes in the GP IIb/IIIa receptor, causing acute and severe thrombocytopenia several hours after drug administration [1,3]. A delayed thrombocytopenia is unexpected, as tirofiban is cleared early from the circulation, with a half-life of 2 h. We postulate that the nature of delayed GP IIb/IIIa inhibitor-induced thrombocytopenia is consistent with a primary immune response where antibodies develop after several days of drug exposure. Such a mechanism was shown for the recently described cases of delayed thrombocytopenia associated with abciximab in which drug-dependent antibodies were demonstrable [2]. However, whilst antibody formation with drug exposure is easily understood, the mechanism of ongoing antibody binding to GP IIb/IIIa is less clear, as it appears to be drug-dependent in vitro. With regards to abciximab, there is evidence of platelet internalization of drug with binding to non-surface expressed GP IIb/IIIa [4], with persisting platelet inhibition well beyond the drug half-life [5]. Whether such a phenomenon occurs with tirofiban is speculative. There is evidence of an increased rate of tirofiban-induced thrombocytopenia in association with other stimuli of platelet activation, presumably due to conformational changes in the GP IIb/IIIa receptor and exposure of neoepitopes [6]. Perhaps platelet activation in vivo allows persistent antibody binding in the absence of drug.

In our institution, a protocol has been developed to identify cases of acute thrombocytopenia by performing a platelet count at 6 and 24 h following tirofiban administration with coronary angioplasty. However, the possibility of delayed thrombocytopenia may justify monitoring platelet counts up to a week following administration of tirofiban. Furthermore, measurement of drug-dependent antibodies is possible days after drug discontinuation and may be useful in confirming the cause of thrombocytopenia.


We would like to thank Sue Evans and Leonie Gaudry for their work.