Is bivalirudin just an expensive heparin?: Table 1

This editorial refers to ‘Bivalirudin, glycoprotein inhibitor, and heparin use, and association with outcomes of primary percutaneous coronary intervention in the United Kingdom’†, by A. Sirker et al. on page 1312. Percutaneous coronary intervention (PCI) in combination with antithrombotic drugs is the most effective therapy for acute coronary syndromes. The most frequently used injectable anticoagulants in contemporary practice are heparin and bivalirudin. Heparin was discovered by McLean and Howell 100 years ago—in 1916.1 The effects of purified heparin in patients were first studied in 1935.2 For the next 80 years heparin remained the most effective and most widely used parenteral anticoagulant. Heparin has certain theoretical disadvantages: it requires antithrombin III to achieve its anticoagulation effect, it has some natural inhibitors, and it may cause a rare but serious heparin-induced thrombocytopenia. The main advantage of heparin is the extensive clinical experience obtained during the last 80 years and its low price and thus wide availability. In the real-life practice of interventional cardiology worldwide, unfractionated heparin (UFH) remains the cornerstone of anticoagulation treatment. To prevent bleeding complications, the heparin dosage should always be calculated individually per patient body weight: the recommended bolus is 60–100 units per kg (a lower dose for elderly patients with low body weight and for elective procedures in stable coronary artery disease; a higher dose for younger patients with acute coronary syndromes). A frequent mistake in real-life practice is the use of an arbitrary UFH dose (e.g. 5000 units or 10 000 units) for all patients, irrespective of body weight. An arbitrary 5000 unit dose will certainly be ineffective for a 95 kg middle-aged smoker, and, on the other hand, an arbitrary 10 000 unit dose may be an extremely dangerous overdose (risk of severe bleeding) for an elderly 55 kg female. Bivalirudin is a synthetic analogue of hirudin. In 1884, Haycraft3 described hirudin as an anticoagulant produced in the saliva of leeches. Direct-acting thrombin inhibitors (hirudin, desirudin, lepirudin, and bivalirudin) have several theoretical advantages over heparin. Bivalirudin use in cardiology dates back 20 years.4 The first randomized study5 comparing heparin and bivalirudin showed no differences in mortality and reinfarction rates and a higher bleeding rate in the heparin group, which was clearly related to an extremely high dosage of heparin (175 units per kg bolus—a dose twice as high as the currently used dosage!!). This study protocol probably inspired many subsequent randomized trials, where the heparin arm usually used a combination with potent glycoprotein (GP) IIb/IIIa inhibitors—and thus was likely to cause more bleeding. In this issue of the journal, the British authors Sirker et al. present data from a large UK registry with a focus on parenteral anticoagulants use.6 This so far largest primary PCI registry did not confirm previously suggested mortality benefits of bivalirudin over heparin and thus (together with the HEAT-PPCI study) questioned the conclusions of the HORIZONS trial (which led to a shift in guidelines and practice in some countries). In the HORIZONS-AMI trial,7 bivalirudin alone was superior to combined therapy with UFH plus a systematic GPIIb/IIIa inhibitor (mostly abciximab). However, the net adverse clinical endpoint (9.2% vs. 12.1%) included major bleeding (4.9% vs. 8.3%). There was no significant difference in the ischaemic endpoints, even though there was a higher incidence of stent thrombosis in the bivalirudin group (1.3% vs. 0.3%). It is not surprising that combination of a potent anticoagulant and a potent antiplatet agent was more harmful than a potent anticoagulant alone. The same problem arises in the design of the EUROMAX trial,8 comparing a single drug strategy (pre-hospital bivalirudin) vs. a combination strategy [heparin with optional (in 69% of patients) GPIIb/IIIa inhibitors]. The primary endpoint [death or non-coronary artery bypass graft (CABG) major bleeding at 30 days] again included bleeding (2.6% vs. 6.0%) and was significantly lower with pre-hospital administration of bivalirudin (5.1% vs. 8.5%). Again, similarly to the HORIZONS-AMI trial, there were no differences in ischaemic endpoints: death (2.9% vs. 3.1%), stent thrombosis (1.6% vs. 0.5%), or re-infarction (1.7% vs. 0.9%). The HEAT-PCI study9 compared bivalirudin vs. UFH with similar rates (15%) of GPIIb/IIIa inhibitors in both arms. The study better represents contemporary practice (restriction of GPIIb/IIIa inhibitors to bail-out situations, the use of novel P2Y12 inhibitors, a radial approach, and implantation of a drug-eluting stent). The primary efficacy endpoint (all-cause mortality, stroke, recurrent infarction, and unplanned target lesion revascularization) was higher in the bivalirudin than in the UFH group (8.7% vs. 5.7%), including an increase in stent thrombosis (3.4% vs. 0.9%), but no significant difference in mortality (5.1% vs. 4.3%). The primary safety outcome (major BARC 3–5 bleeding) was 3.5% in the bivalirudin group vs. 3.1% in the UFH group. Thus, the entire treatment benefit of bivalirudin demonstrated in HORIZONS-AMI and EUROMAX trials was caused by the study design (low use of GPIIb/IIIa inhibitors in the bivalirudin arm) and cannot answer the question of whether bivalirudin is superior to heparin or vice versa. The results of the HEAT-PCI study suggest that heparin may even be superior to bivalirudin when the same rate of GPIIb/IIIa inhibition is used. Furthermore, a recent meta-analysis of 11 randomized trials10 demonstrated the increased risk of definite stent thrombosis after bivalirudin compared with heparin [relative risk (RR) 1.73; 95% confidence interval (CI) 1.24–2.40; P < 0.001]. Another recent meta-analysis11 of trials in patients with ST-segment elevation myocardial infarction (STEMI) concluded that bivalirudin, as compared with heparin, increases the risk of stent thrombosis and ischaemia-driven repeat revascularization at 30 days. There is no strong evidence that bivalirudin significantly reduces major bleeding at 30 days. Bivalirudin does not have an effect on all-cause mortality at 30 days. The data from these two meta-analyses plus the HEAT-PPCI study, plus the UK registry in this issue, form a strong base to reject the theory that bivalirudin is superior to heparin. Table 1 summarizes all randomized trials comparing heparin and bivalirudin in the interventional cardiology setting. Trials are divided into two groups: those with a design favouring the bivalirudin arm (by planning a more aggressive antithrombotic strategy in the control arm) and those with a balanced design (the same conditions for the bivalirudin and heparin arms). The most recent study12 randomized patients with acute myocardial infarction into three groups: bivalirudin alone, heparin alone, and heparin plus tirofiban. This study demonstrated similar outcomes after bivalirudin as after heparin, while an increased number of events (mainly bleeding) was observed after combined heparin + tirofiban. This confirms the above-mentioned critical view on many previous trials, where bivalirudin was the winner due to the study design. Randomized trials comparing heparin and bivalirudin in the interventional cardiology setting Red colour: trials with higher use of GPIIbIIIa inhibitors in the control arm. Green colour: trials with comparable use of GPIIbIIIa inhibitors in both arms. ACS, acute coronatry syndrome; ACT, activated clotting time; AMI, acute myocardial infaction; CAD, coronary artery disease; DM, diabetes mellitus; NSTE, non-ST-segment elevation; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction. Randomized trials comparing heparin and bivalirudin in the interventional cardiology setting Red colour: trials with higher use of GPIIbIIIa inhibitors in the control arm. Green colour: trials with comparable use of GPIIbIIIa inhibitors in both arms. ACS, acute coronatry syndrome; ACT, activated clotting time; AMI, acute myocardial infaction; CAD, coronary artery disease; DM, diabetes mellitus; NSTE, non-ST-segment elevation; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction. In this context it is interesting to look back to 1994, when Biogen stopped development of bivalirudin after some disappointing study results, and sold the licence to The Medicines Company in 1997. In 1998, the US Food and Drug Administration advisory panel voted against recommending approval of bivalirudin for use in patients with unstable angina who are undergoing angioplasty, because it was equivalent but not superior to heparin in preventing angioplasty complications. Panel members expressed doubt about the company's claim that bivalirudin was safer (less major bleeding), because the heparin dose used in the phase III studies was much higher than that used routinely in current practice. Furthermore, other direct thrombin inhibitors (Revasc, desirudin; Refludan, lepirudin; and Novastan, argatroban) did not really penetrate into cardiology practice. Jeffrey Swarz (Credit Suisse First Boston) commented in 1998 that the key problem for these products is price, because heparin works and it is cheap. This comment is still valid after 17 years with the current large price differences (bivalirudin is 10–50 times more expensive than heparin). Conflict of interest: none declared