Comparing the safety of ticagrelor versus clopidogrel: insights from the FDA reports

Race and gender

The PLATO trial enrolled over a thousand (1081) patients of Chinese or Japanese descent [FDA, 2011, 2012]. Importantly, Japanese patients experienced a 40% greater exposure to ticagrelor versus Caucasians, which was shown in an 8-day phase I study [both for maximum concentration (Cmax) and for under the curve (AUC)] [FDA, 2012]. While there are insufficient safety data in this subgroup, Japanese patients may have an increased risk for overdosing on ticagrelor compared with their white counterparts. However, data are lacking to confirm that higher exposure to ticagrelor in Japanese patients leads to an increased risk of hemorrhagic events. Moreover, women on ticagrelor tended to have more adverse events (AEs), including serious AEs (SAEs), increased discontinuation because of AEs and an increased risk of death compared with men, also suggesting potential overdosing [FDA, 2012]. Whether these differences were statistically significant was not mentioned in the FDA complete response review (CRR) [FDA, 2011].

Bleeding

The primary safety concern with ticagrelor in the PLATO trial is bleeding. The FDA CRR indicates that nonprocedural bleeding events were significantly increased after ticagrelor versus clopidogrel [major + minor + minimal: 24.3% versus 15.9%, HR 1.62, 95% CI 1.51–1.74); major + minor: 5.9% versus 4.3%, HR 1.39, 95% CI 1.21–1.60; major: 3.1% versus 2.3%, HR 1.31, 95% CI 1.08–1.60) with a nonsignificant 9% increase in major life-threatening or fatal bleeds (1.3% versus 1.2%, HR 1.09, 95% CI 0.82–1.44) and fatal bleeds (0.2% versus 0.2%, HR 1.09, 95% CI 0.50–2.38) respectively [FDA, 2011]. Thus, ticagrelor significantly increases major, minor and minimal nonprocedural bleeds compared with clopidogrel. Furthermore, major bleeds were predominately coronary artery bypass graft (CABG)-related bleeding complications (~67%) [FDA, 2011]. However, in PLATO, the reporting of (CABG) bleeds and procedural hemorrhages was set up in ticagrelor’s favor. Indeed, the reporting of procedural bleeds was set up to be related to the procedure rather than the study drug. The protocol stated that procedural bleeding should not be reported as an AE if it is expected during the procedure, yet there was no guidance on the form about what is ‘expected’ during the procedure [FDA, 2011]. Furthermore, on the AE form, the checkbox for ‘Bleed is related to a procedure and does not represent an adverse event’ did not allow the severity of the bleed (minimal, minor or major) to be recorded and the remainder of the AE form was not to be completed. However, the reporting of CABG-related bleeds was set up to document as much as possible (i.e. the electronic case report form system automatically created a bleed event form for all CABG forms) as the ‘CABG-related’ checkbox also had minimal, minor and major bleeding severity checkboxes (unlike the reporting for procedural bleeds) [FDA, 2011]. Moreover, the investigator was to complete the rest of the AE form for CABG-related bleeds, which was not required for procedural-related bleeds [FDA, 2011]. As ticagrelor has a faster offset, one would expect fewer CABG bleeds but more procedural-related bleeds due to ticagrelor’s higher inhibition of platelet activity compared with clopidogrel. Thus, it seems that the PLATO investigators stacked the deck of cards in ticagrelor’s favor (i.e. reporting more CABG bleeds but fewer procedural bleeds). Despite this ploy, the risk of CABG bleeding was increased in patients receiving ticagrelor compared with clopidogrel-treated patients in those who did not wait until day 5 after stopping treatment, which is recommended in ACS guidelines [FDA, 2011; Levine et al. 2011]. Thus, the antiplatelet effects of ticagrelor last significantly longer than predicted by its pharmacokinetics/pharmacodynamics. In addition, the purported ‘faster’ offset of ticagrelor (compared with clopidogrel) is not clinically relevant as it does not translate into fewer bleeding events. There were statistically more overall bleeds (PLATO-defined major, life-threatening or fatal, minor and minimal) in the ticagrelor arm [FDA, 2011]. In summary, in patients treated with ticagrelor, clinicians should consider delaying CABG and other major surgeries until 5 days after cessation of therapy.

Strokes and embolic events

There were more intracranial bleeds on ticagrelor compared with clopidogrel [27 (0.3%) versus 14 (0.2%), relative risk (RR) 1.90, p value not stated; one extra intracranial bleed for every 1000 patients treated] and more hemorrhagic strokes (26 versus 16, RR 1.63, p value not given) with 11 (0.12%) fatal hemorrhagic strokes on ticagrelor versus only 1 (0.0%) on clopidogrel (i.e. there was a 10-fold increase in fatal hemorrhagic strokes with ticagrelor versus clopidogrel; p = 0.02) [FDA, 2011]. Furthermore, for every 10 all-cause deaths avoided with ticagrelor, there was one extra hemorrhagic stroke death [FDA, 2011]. A higher degree of platelet inhibition with ticagrelor may precipitate a clinical hemorrhagic event in someone with a subclinical stroke compared with someone on clopidogrel. Supporting this theory was the significant increase in intracranial hemorrhages (ICHs) on ticagrelor compared with clopidogrel (27 versus 14) as well as fatal ICH [11 (0.1%) versus 2 (0.0%), RR 5.47; 1 extra fatal ICH for every 1000 patients treated with ticagrelor]. Moreover, peripheral vascular disease was increased by 17% (RR 1.17, 155 versus 132), thromboembolic disease (arterial and venous) was increased by 24% (RR 1.24, 84 versus 68), any venous thrombotic event was increased by 11% (RR 1.11, 72 versus 65), deep vein thrombosis was increased by 11% (RR 1.11, 72 versus 65) and pulmonary embolism was increased by 46% (35 versus 24) with ticagrelor compared with clopidogrel (Table 1) [FDA, 2011]. A higher platelet inhibition may lead to clots that are more brittle and thus more likely to embolize on ticagrelor. Another explanation purported by the FDA document was that ticagrelor may increase the risk for plaque rupture and subsequent embolization further downstream in the vasculature. In summary, a higher antiplatelet potency and ‘faster’ antiplatelet offset with ticagrelor does not seem to give it a clinical advantage with regard to reduced embolic events or reduced hemorrhagic events.

OPEN IN VIEWER

Table 1.

Embolic/vascular events: ticagrelor compared with clopidogrel [FDA, 2011].

Event	Ticagrelor	Clopidogrel	RR
Amaurosis fugax	7	3	2.33
Stroke	138	111	1.24
Intracranial hemorrhage	28	17	1.65
Stroke/TIA	178	152	1.17
Cerebrovascular and carotid disease	214	190	1.13
Thromboembolic disease (arterial and venous)	84	68	1.24
Peripheral vascular disease	155	132	1.17
Necrotic limb/amputation	4	1	4
Venous event	72	65	1.11
DVT	49	43	1.14
PE	35	24	1.46

DVT, deep vein thrombosis; PE, pulmonary embolism; RR, relative risk; TIA, transient ischemic attack.

Dyspnea

Shortness of breath or dyspnea was reported more frequently with ticagrelor compared with clopidogrel (14.49% versus 8.78%, RR 1.6, p value not stated), with one-third of patients experiencing dyspnea throughout study termination [FDA, 2011]. Thus, one out of three patients experiencing dyspnea on ticagrelor will continue to have this serious adverse event for a prolonged period of time. Moreover, dyspnea leading to discontinuation was nine times higher with ticagrelor versus clopidogrel (0.9% versus 0.1%, RR and p value not given). Discontinuation for dyspnea (from sponsor) indicated an almost eightfold increase with ticagrelor versus clopidogrel [77 (0.8%) versus 10 (0.1%), RR 7.66] [FDA, 2011]. Thus, the reported side effect of ‘dyspnea’ with ticagrelor seems to be more clinically relevant than originally stated [Wallentin et al. 2009; FDA, 2011]. While a pulmonary function substudy was performed, flaws in its design, conduct and analysis prevented any ascertainment of conclusions. High baseline smoking status as well as lack of the preferential inclusion of individuals who experienced dyspnea on ticagrelor may have suppressed a difference, even if one had existed [FDA, 2011].

Gynecomastia

While the absolute rate of gynecomastia was low (approximately 3 per 1000 men at 1 year), ticagrelor significantly increased gynecomastia 5.3 fold compared with clopidogrel [15 (0.16%) versus 3 (0.03%), RR > 5.0]. Thus, in patients experiencing or having a history of gynecomastia, clopidogrel may be preferred.

Arrhythmias and bradycardia

The DISPERSE 2 trial, a phase II trial comparing two doses of ticagrelor versus clopidogrel in 990 patients, indicated a higher frequency of arrhythmias with ticagrelor [FDA, 2011]. In PLATO, while there were more atrial arrhythmias and ventricular pauses with ticagrelor versus clopidogrel, a somewhat favorable effect was seen for sudden death and ventricular arrhythmias [FDA, 2011]. However, PLATO excluded patients at increased risk of bradycardic events as ticagrelor was shown to significantly increase this side effect compared with clopidogrel in the DISPERSE 2 trial [FDA, 2011]. Thus, patients at risk of a bradycardia should not be placed on ticagrelor.

History of hepatic disorder

Approximately 400 patients with a ‘history of baseline hepatic disorder’ were enrolled in the PLATO trial [FDA, 2011]. In these individuals, compared with clopidogrel, ticagrelor significantly increased death by over threefold (3.1% versus 0.9%), and frequency of SAEs (20.4% versus 16.6%) and AEs (84.2% versus 81.1%). Moreover, ticagrelor increased major bleeds compared with clopidogrel in these individuals (11.2% versus 8.7%) [FDA, 2011]. In summary, clinicians should use caution when prescribing ticagrelor in patients with a history of hepatic disorder.

Renal insufficiency

Compared with clopidogrel, patients receiving ticagrelor with a baseline estimated glomerular filtration rate (eGFR) of less than 30 ml/min had an increased risk of major bleeding complications [23 (19%) versus 16 (11.3%)], death [31 (26.5%) versus 34 (23.4%)] and renal failure [12 (13.6%) versus 5 (5.4%)] respectively. This contrasts with a recent publication by James and colleagues, indicating that a cohort of 214 patients with calculated creatinine clearance less than 30 ml/min showed similar results for the primary composite endpoint, 28.9% per 12 months (27 patients) for ticagrelor versus 39.0% (39 patients) for clopidogrel (HR 0.77; 95% CI 0.49–1.30), and for mortality, 23.4% (21 patients) versus 29.6% (29 patients) (HR 0.77; 95% CI 0.47–1.44) [FDA, 2011; James et al. 2010]. Thus, in patients with a baseline eGFR less than 30 ml/min, ticagrelor may have a net clinical harm compared with clopidogrel.

Drug interactions

Ticagrelor is a cytochrome P450 (CYP)-3A4 substrate and a relatively weak CYP3A4 inhibitor, increasing the exposure of simvastatin (50%, up to two- to threefold in some individuals), atorvastatin (36%), lovastatin (similar to simvastatin) and estradiol (30% in Cmax and 20% in AUC) [FDA, 2011]. CYP3A inducers such as rifampin, dexamethasone, phenytoin, carbamazepine and phenobarbital increase ticagrelor’s clearance by 110%. Thus, patients on CYP3A inducers may have less benefit on ticagrelor. In summary, CYP3A4 inhibitors, inducers and substrates need to be given with caution with concomitant ticagrelor administration. Moreover, ticagrelor inhibits the p-glycoprotein (PgP) efflux pump and can significantly increase digoxin levels (75% increase in Cmax, 28% increase in AUC; mean digoxin trough levels increased by 30%, with some individual maximum increases of up to twofold). Thus, in patients on ticagrelor, digoxin concentrations should be monitored closely. Furthermore, concomitant use of ticagrelor and other PgP inhibitors such as verapamil, quinidine and cyclosporine should be used with caution.

Despite the favorable efficacy of ticagrelor versus clopidogrel, there are a plethora of subgroups and clinical scenarios in which the safety of ticagrelor may outweigh its benefit (Table 2). When prescribing ticagrelor, clinicians should be fully aware of these potential harmful scenarios, and have an action plan in case adverse events emerge. While ticagrelor has very limited (if any) ‘real-world’ experience, the progressive years of its inclusion in the ACS preventive cocktail should allow us to better gleam its overall efficacy and safety profile.

OPEN IN VIEWER

Table 2.

Scenarios in which ticagrelor safety may outweigh benefit [FDA, 2011; EMA, 2010].

CYP3A4 substrates, inhibitors and inducers

PgP substrates/inhibitors – digoxin, cyclosporine and verapamil

Women and Japanese

Baseline GFR < 30 ml/min

History of cerebrovascular disease

History of a thromboembolic event

History of ICH

History of Hepatic disease

Patients at risk of bradycardia

CYP3A4, cytochrome P450 3A4; GFR, glomerular filtration rate; ICH, intracranial hemorrhage; PgP, P glycoprotein.