Retrospective analysis of arterial occlusive events in the PACE trial by an independent adjudication committee

Abstract

Background

The phase 2 PACE (Ponatinib Ph+ ALL and CML Evaluation) trial of ponatinib showed robust long-term benefit in relapsed Philadelphia chromosome-positive (Ph+) leukemia; arterial occlusive events (AOEs) occurred in ≥ 25% of patients based on investigator reporting. However, AOE rates vary depending on the definitions and reporting approach used.

Methods

To better understand clinically relevant AOEs with ponatinib, an independent cardiovascular adjudication committee reviewed 5-year AOE data from the PACE trial according to a charter-defined process and standardized event definitions.

Results

A total of 449 patients with chronic myeloid leukemia (CML) or Ph+ acute lymphoblastic leukemia (ALL) received ponatinib (median age 59 y; 47% female; 93% ≥ 2 prior tyrosine kinase inhibitors (TKIs); median follow-up, 37.3 months). The adjudicated AOE rate (17%) was lower than the non-adjudicated rate (i.e., rate before adjudication; 25%). The only adjudicated AOE in > 2% of patients was peripheral arterial occlusive disease (4%). Exposure-adjusted incidence of newly occurring adjudicated AOEs decreased over time. Patients with multiple baseline cardiovascular risk factors had higher adjudicated AOE rates than those without risk factors.

Conclusions

This independent adjudication study identified lower AOE rates than previously reported, suggesting earlier overestimation that may inaccurately reflect AOE risk with ponatinib. This trial was registered under ClinicalTrials.gov identifier NCT01207440 on September 23, 2010 (https://clinicaltrials.gov/ct2/show/NCT01207440).

Background

Ponatinib, a pan-BCR::ABL1 inhibitor, is an orally active third-generation tyrosine kinase inhibitor (TKI) designed to potently inhibit BCR::ABL1 with or without any point mutation, including BCR::ABL1^T315I [1]. In the pivotal phase 2 PACE (Ponatinib Ph+ ALL and CML Evaluation) trial, ponatinib demonstrated robust clinical activity with rapid, deep, and long-term responses, progression-free survival (PFS), and overall survival in patients with chronic-phase chronic myeloid leukemia (CP-CML), ≥ 90% of whom had failed treatment with ≥ 2 TKIs, regardless of the presence or absence of BCR::ABL1 mutations, including T315I [2, 3]. The 5-year results of the PACE trial confirmed the durability of these responses with a 5-year overall survival rate of 73% for CP-CML [3]. However, arterial occlusive events (AOEs) were reported by investigators in 25% in the overall population (serious AOEs, 20%) and 31% in the CP-CML population (serious AOEs, 26%) in the 5-year follow-up [3]. The exposure-adjusted incidence of newly occurring AOEs decreased from year 1 (15.8 patients with events per 100 patient-years in the total population) to year 5 (3.9 per 100 patient-years) [3]. The incidence of AOEs associated with ponatinib use has varied widely in subsequent reports. Two retrospective studies have reported an absence or very low incidence (6%) of AOEs [4, 5]. Other real-world studies have reported AOE rates ranging from 18 to 26% [6, 7]. Multiple factors may contribute to variability in reported AOE rates, including differences in patient populations, as well as differences in the clinical definitions used to identify and categorize vascular occlusive events. One of the most important factors is the lack of a standardized approach for defining and capturing AOEs with BCR::ABL1 TKIs.

The AOE incidence rate reported for PACE was based on a list of approximately 400 Medical Dictionary for Regulatory Activities (MedDRA) preferred terms developed by the sponsor. However, differences in the preferred terms used to define AOEs led to variability in AOE incidence rates. Some preferred terms included in the AOE analysis of PACE are highly sensitive for identification of potential AOEs but may not themselves indicate the occurrence of arterial occlusions, frequently including symptoms or descriptions rather than events; these include chest pain, cold hands, dysarthria, and poor peripheral circulation. This approach to characterize AOEs based on adverse event terms results in broadly capturing non-specific symptoms that may be associated with AOE rather than true AOEs and may thus overestimate the incidence of clinically meaningful events.

A clear understanding of clinically relevant AOE risk is imperative when characterizing the benefit-risk profile of ponatinib. Patients with CP-CML who become resistant to a second-generation BCR::ABL1 TKI, either with or without a BCR::ABL1 gene mutation, generally experience low response rates and poor survival if treated with another second-generation TKI [8, 9]. Importantly, ponatinib is the only currently available TKI effective in patients with the BCR::ABL1^T315I mutation [3]. Therefore, the potential for improved survival and duration of response on ponatinib may outweigh the risk of AOEs [8, 9]. However, the lack of clear data regarding clinically meaningful AOEs has led to confusion about how to optimally use ponatinib to treat relapsed/refractory CML and Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukemia (ALL) and, in some instances, avoidance in patients who could potentially benefit. To provide a more accurate characterization of AOE incidence with ponatinib, an independent adjudication committee of experts was convened to retrospectively adjudicate all AOE reports in the PACE trial in a standardized, rigorous manner.

Methods

PACE trial design

The phase 2 PACE trial (ClinicalTrials.gov identifier: NCT01207440) enrolled adults with CML or Ph+ ALL whose disease was resistant or intolerant to dasatinib or nilotinib, or who had the BCR::ABL1^T315I mutation regardless of prior TKI use [3]. All patients received ponatinib at a starting dose of 45 mg once daily (qd); dose reductions to 30 or 15 mg qd were applied per protocol (Table 1) to manage adverse events (AEs), or implemented proactively following recommendations from the sponsor in October 2013 in response to AOEs emerging as notable AEs. The trial has been completed; detailed methods are published [2, 3].

Table 1 Dose reduction recommendations (as of 2013)

Adjudication methods

All activities related to the adjudication of AOEs were conducted by ACI Clinical (Bala Cynwyd, PA), including the identification of an independent adjudication committee. ACI Clinical is a clinical research organization with expertise in Endpoint Adjudication and Data Monitoring Committees to support safety decisions around clinical development programs. ACI Clinical was contracted by the sponsor; adjudication activities were not part of the PACE trial.

Identification of AEs for adjudication

To ensure all relevant potential events were captured, the PACE AE dataset (449 patients with 12,224 AE records; extraction date: May 9, 2018) was searched using a comprehensive set of 604 preferred terms potentially relevant to AOEs that was developed by the sponsor (Table 2). This search strategy, which was more comprehensive than that used in initial analyses of the PACE trial, identified 181 patients and 455 AE records for adjudication (Fig. 1A). In addition, all patient deaths not attributable to disease progression by the clinical investigator were reviewed by the chair of the adjudication committee (described below) for identification of potential fatal AOEs. The adjudication committee identified 45 fatal events for review. In total, 202 patients and 490 events were submitted to the independent adjudication committee for review (Fig. 1A).

Table 2 List of 604 preferred terms used to identify AEs for adjudication

Fig. 1

CONSORT flow diagram and process for adjudication of arterial occlusive events (AOEs). A CONSORT diagram: Identification of AOEs for review by the adjudication committee. B Adjudication process flow charts. AE adverse event, AC adjudication committee, AIM Applied Clinical Intelligence Information Management System, MedDRA Medical Dictionary for Regulatory Activities, PACE Ponatinib Ph+ ALL and CML Evaluation, PE pulmonary embolism, PVD peripheral vascular disease, VTE venous thromboembolism. ^aThe Adjudication Committee also reviewed any events included in the Cardiac Failure Standard MedDRA Query (SMQ) to determine whether any heart failure events were AOEs. ^bAOEs evaluated on the left panel excluded events evaluated in the right panel (stroke, DVT, and PE). ^cPer the charter, panel meetings were convened to discuss events for which a decision was not reached via independent voting. The quorum for panel meeting attendance was dependent on the type of event(s) to be discussed (i.e., cardiologist, neurologist, or vascular specialist)

An individual case package containing all available clinical information (including medical history) was created for each event and provided to the adjudication committee members for their review. If a patient experienced more than 1 event within 48 h, these events were adjudicated as potentially representing a single clinical event, unless the case evidence suggested they were independent events. Individual events occurring > 48 h apart were adjudicated as independent events. All data were from the clinical trial database that was in SAS format and structured in conformance to CDISC SDTM format; no other source material was available.

Adjudication procedure

An adjudication committee of academic research clinicians who are highly experienced in adjudication activities in cardiovascular trials was appointed by ACI Clinical. The adjudication committee of 5 independent academic experts (3 cardiologists, 1 vascular medicine specialist, and 1 vascular neurologist) retrospectively adjudicated suspected cases of arterial occlusive events in the PACE study. The committee followed a predefined process outlined in the adjudication charter developed by ACI clinical. The charter defined the responsibilities of the adjudication committee and the adjudication endpoints using established definitions developed by the 2014 American College of Cardiology (ACC)/American Heart Association (AHA) guideline [10], and the definitions for cardiovascular and stroke outcomes developed by the Standardized Data Collection for Cardiovascular Trials Initiative (SCTI) and the US Food and Drug Administration [11, 12]. All suspected AOEs identified in the PT search were assessed using the charter definitions (Table 3) for myocardial infarction; heart failure if attributed to an AOE, which may include coronary artery disease, arterial hypertension, cardiomyopathy, or myocardial infarction; hospitalization for unstable angina; stroke and other cerebrovascular events; and peripheral vascular disease. Any events meeting the criteria of these endpoints were considered adjudicated AOEs. Specific criteria were required (e.g., revascularization, change in cardiac biomarkers, diagnostic evidence as shown by computerized tomography scan, magnetic resonance imaging, etc.) to determine the presence of a clinical endpoint. The adjudication committee members were blind to ponatinib dose at the time of the event, whether dose modifications were made, and the investigator’s opinion on AE causality.

Table 3 Adjudication committee prespecified definitions of events

During the adjudication process, the committee reviewed all potential AOEs, as well as any AEs identified in a Cardiac Failure Standard MedDRA Query (SMQ), to determine whether any heart failure events were AOEs. Two members of the adjudication committee independently evaluated whether an individual case met the prespecified event definitions (Fig. 1B). If agreement between 2 members was not reached for cases of AOEs or heart failure, the case was reviewed by a third cardiologist adjudication committee member; if agreement was not reached with 3 votes, the case was reviewed at a panel meeting. If agreement was not reached for cases of stroke, deep vein thrombosis, pulmonary embolism, and peripheral vascular disease, the case was discussed at a panel meeting with the appropriate neurologist and/or vascular specialist member(s). All fatal events were decided by consensus of adjudicators.

Events that met one of the charter-defined endpoint definitions were further categorized depending on the event type (e.g., myocardial infarction, peripheral arterial occlusive disease, deep vein thrombosis, etc.). Non-adjudicated AOEs that were recorded as symptoms (e.g., "non-cardiac chest pain" or "claudication") with a low severity level and no accompanying changes in medication or hospitalization were adjudicated to not be AOEs unless they had an anatomic diagnosis provided (e.g., "severe superficial femoral artery stenosis"). If the term "infarction" was provided for stroke events, the adjudicators categorized the event as ischemic stroke. Revascularization was not always clearly reported by investigators.

Statistics

Exposure-adjusted AOE rates were calculated as: (number of first events in interval)/(total exposure for interval in patient-years) × 100. The relative risk of serious AOEs was analyzed by baseline risk category in patients from the safety population for whom all baseline risk categories were available. Risk categories included commonly recognized cardiovascular risk factors for which data were collected (arterial hypertension, hypercholesterolemia, diabetes mellitus, and obesity), and history of heart disease (non-ischemic or ischemic).

Results

Patient disposition and baseline characteristics

Patient disposition and baseline characteristics in the PACE trial have been published [2, 3]. A total of 449 patients, including 270 CP-CML patients, 85 accelerated-phase (AP) CML patients, 62 blast-phase (BP) CML patients, and 32 Ph+ ALL patients, were enrolled between September 2010 and October 2011. Baseline characteristics are summarized in Table 4. Among all 449 patients, the median age was 59 years and 53% of patients were male. Most (93%) patients had received 2 or more prior TKIs. At baseline, 53% of patients had arterial hypertension, 49% had hypercholesterolemia, and 24% had BMI ≥ 30 kg/m². Forty-three percent of patients had a baseline history of non-ischemic cardiac disease, and 23% had a history of ischemic cardiovascular disease. Safety data reviewed by the adjudication committee reflect data collected as of February 6, 2017, with median follow-up of 37.3 months for all patients and 56.8 months (range 0.1–73.1 months) for CP-CML patients.

Table 4 Baseline characteristics and disposition at end-of-study³

Adjudication results

Rates of adjudicated AOEs were lower than rates of non-adjudicated AOEs (Fig. 2A). Overall, 17% (78/449) of patients had adjudicated AOEs compared with 25% (111/449) with non-adjudicated AOEs. Most patients with serious AOEs were adjudicated as having serious AOEs (20% [90/449] non-adjudicated vs. 16% [74/449] adjudicated). Most (95% [74/78]) patients with adjudicated AOEs had serious AOEs. In CP-CML patients, rates of adjudicated AOEs (21% [57/270]) were also lower than rates of non-adjudicated AOEs (31% [84/270]); 95% [54/57] of CP-CML patients with adjudicated AOEs had serious AOEs. The rates of AOEs by AOE type (i.e., cardiovascular, cerebrovascular, and peripheral vascular) are presented for all patients in Table 5 and for CP-CML patients in Table 6.

Fig. 2

Arterial occlusive event (AOE) rates with ponatinib. A Rates of non-adjudicated and adjudicated AOEs. B Exposure-adjusted incidence of newly occurring arterial occlusive events (AOEs) by year (all patients). Later intervals excluded patients with prior events. Non-adjudicated values were published previously [3]. C Incidence of AOEs (adjudicated) by number of baseline risk factors (all patients). Risk factors included arterial hypertension, hypercholesterolemia, obesity, diabetes mellitus, non-ischemic cardiac disease, and ischemic disease. D Overall survival (OS) in chronic-phase chronic myeloid leukemia (CP-CML) patients with and without AOEs

Table 5 Rates of non-adjudicated and adjudicated AOEs by type in the total population (n = 449)

Table 6 Rates of AOEs non-adjudicated and adjudicated AOEs in CP-CML patients (n = 270)

The most common non-adjudicated and adjudicated AOEs and serious AOEs are summarized in Table 7. The most common (> 2%) non-adjudicated AOEs were angina pectoris (6%; 28/449), peripheral arterial occlusive disease (5%; 22/449), MI (4%; 18/449), coronary artery disease (3% [14/449]). The only adjudicated AOE reported in > 2% of patients was peripheral arterial occlusive disease (4% [16/449]). Non-adjudicated AOEs that were most commonly adjudicated as not AOEs were angina pectoris, non-cardiac chest pain, and chest pain, as these events were often recorded as symptoms (e.g., "non-cardiac chest pain" or "claudication") or presumptive diagnoses with a low severity level and no accompanying changes in medication or hospitalization.

Table 7 Arterial occlusive events (AOEs) in ≥ 2.0% of patients (n = 449)

The exposure-adjusted incidence of adjudicated AOEs (8.9 patients with events per 100 patient-years) and serious AOEs (8.4 patients with events per 100 patient-years) was lower than the exposure-adjusted incidence of non-adjudicated AOEs (11.3 and 9.2 per 100 patient-years, respectively). The exposure-adjusted incidence of newly occurring AOEs decreased over time (Fig. 2B). The median time to onset of the first adjudicated AOE was 14.1 months (range: 0.1 to 49.5; Table 8).

Table 8 Time to onset of adjudicated AOEs

Resolution of AOEs, dose modifications, and discontinuations

Among the 78 patients with an adjudicated AOE, events resolved in 51 patients. Among 43 patients with just one AOE, 74% (32/43) had resolution of the event; 35 patients had multiple AOEs recorded, with 54% (19/35) patients having resolution of all the events. Most patients continued ponatinib after the AOE, including 36 patients (46%) who continued ponatinib without dose modification and 27 patients (35%) who had their doses reduced and/or interrupted after the event (Table 9). Seven patients (9%) discontinued ponatinib due to an adjudicated AOE. Rates of dose modifications following AOEs are summarized in Table 9.

Table 9 Ponatinib dose modifications following non-adjudicated and adjudicated arterial occlusive events (AOEs)^a

Risk factor analysis

The most common baseline risk factors in patients who developed an AOE were arterial hypertension and hypercholesterolemia (Table 10). Patients with adjudicated AOEs also had higher rates of concomitant use of antihypertensive medications, platelet aggregation inhibitor medications, and anti-diabetic agents compared with patients who did not have AOEs (Table 11).

Table 10 Prevalence of baseline risk factors by adjudicated AOE and serious AOE status

Table 11 Concomitant medication use by adjudicated AOE and serious AOE status

The incidence of adjudicated AOEs by number of baseline risk factors (including arterial hypertension, hypercholesterolemia, obesity, diabetes mellitus, non-ischemic cardiac disease, and ischemic disease) is shown in Fig. 2C. The rate of adjudicated AOEs was 13% (24/189) among patients with 1–2 risk factors, and 29% (52/180) among patients with 3 or more risk factors. Of the 80 patients without any risk factors at baseline, only 2 (3%) had an AOE.

Fatal AOEs

Separate adjudication of deaths revealed that 11 adjudicated AOEs were associated with death. These included 2 cases of cardiac arrest and 1 each of the following: bradycardic arrest, cardiac failure, intracranial hemorrhage, worsening of congestive heart failure, superior mesenteric artery occlusion, hemorrhagic cerebral infarction, congestive heart failure, ischemic stroke, and acute anterior myocardial infarction. Nine of the 11 patients with AOEs associated with death had a history of cardiovascular events and/or cardiovascular risk factors recorded at baseline (Table 12). The long-term survival of patients with adjudicated AOEs was similar to survival of patients without AOEs (Fig. 2D).

Table 12 Fatal AOEs and patient baseline characteristics

Discussion

In this study, adjudication of AOEs by an independent committee of experts allowed for a clinically meaningful description of AOEs associated with ponatinib, which can help to inform health care providers and patients of safety risks in an accurate and objective manner. The search that identified potential AOEs for adjudication was broader (based on 604 MedDRA terms related to vascular ischemia or thrombosis) than that initially used to calculate non-adjudicated AOE rates in the PACE trial (400 MedDRA terms) [3]. Based on 5-year follow-up of the PACE trial, the adjudicated AOE rate (17%) was lower than the non-adjudicated AOE rate (25%) [3]. Although the majority of adjudicated AOEs were serious, 81% of patients with AOEs continued on ponatinib (35% with dose modifications), the benefit of the drug was felt to outweigh the risk of the AOEs. Although vascular occlusive events were rarely reported during the initial development of second-generation BCR::ABL1 TKIs, a meta-analysis found that these events occurred in 5.9% of patients with CML treated with these agents, including bosutinib, dasatinib, nilotinib, and ponatinib [13]. In another review of prospective trials of patients treated with TKIs, including imatinib, nilotinib, dasatinib, and ponatinib, overall incidence of CV events was 45% (range, 41–63%) [14]. Accordingly, a high level of vigilance is indicated to recognize this potential complication of TKI therapy.

Notably, although concern existed around the potential for increasing AOE rates with long-term dosing, as seen with AEs related to other TKIs [15,16,17,18] the exposure-adjusted incidence of newly occurring adjudicated AOEs decreased over time on ponatinib, suggesting that the toxicity of ponatinib may not increase with longer treatment duration.

Patients with adjudicated AOEs were more likely to have multiple baseline cardiovascular risk factors (e.g., ischemic cardiac disease, arterial hypertension, hypercholesterolemia, and diabetes mellitus), and only 2 patients had an adjudicated AOE without any cardiovascular risk factors. These observations align with those of previous studies [6, 19]. It is important to identify and manage cardiovascular risk factors before and during therapy with ponatinib or other TKIs [20,21,22]. In PACE, 80% of CP-CML patients were resistant to dasatinib or nilotinib, and 24% had the BCR::ABL1^T315I resistance mutation [3]. Among CP-CML patients, estimated 5-year PFS and OS rates were 53% and 73%, respectively [3]. Data for overall survival in patients with and without adjudicated AOEs suggest that the risk of AOE-related death did not substantially impact survival, with disease-related death being the main driver of the OS curve. This underscores the need for providers to fully understand the therapeutic profile of ponatinib and consider its use when the potential benefits outweigh the risks for a given patient.

This study reinforces the importance of proper assessment of cardiovascular AEs to ensure accurate estimation of cardiovascular risk. The conventional processes of AE reporting and causality assessment may need to be re-assessed to avoid pitfalls associated with over- or under-reporting of AOEs, both of which may adversely affect patient care [23, 24]. Formal adjudication of events is a mainstay for development programs in other therapeutic areas such as diabetes mellitus [25, 26] and cardiology. A better understanding of the AOE risk associated with TKI therapy is a prime example of where formal adjudication is critical because accurate knowledge of risks is crucial before prescribing any TKI. The potential benefits of effective BCR::ABL1 TKI treatment, even with accompanying AEs, may outweigh the potential risks of progression-related mortality in patients with CP-CML and Ph+ ALL receiving second- or third-line therapy. This is particularly true for patients such as those with the BCR::ABL1^T315I mutation who may have limited treatment options [27]. Understanding the true incidence of the most significant events is a central element in properly assessing the benefit-risk ratio of an intervention. All later-generation TKIs are associated with risk of cardiovascular AEs [28], and the results of the formal adjudication process suggest the risk of these events with ponatinib may not be dissimilar to the event rates seen with some second-generation BCR::ABL1 TKIs [16,17,18].

A noteworthy finding in our analysis is that the exposure-adjusted incidence of newly occurring adjudicated AOEs decreased over time on ponatinib. These results are reassuring that the rate of new AOEs may not increase with longer duration of ponatinib treatment. Furthermore, patients with positively adjudicated AOEs were much more likely to have baseline cardiovascular risk factors (e.g., arterial hypertension, hypercholesterolemia, diabetes mellitus) or established cardiovascular disease; of those patients without any cardiovascular risk factors only 2 had a subsequent AOE. These results may provide clinical guidance with respect to the approach to use of ponatinib in patients at risk for an AOE. The ongoing phase 2 OPTIC trial (ClinicalTrials.gov Identifier: NCT02467270) is using a response-based dose reduction protocol approach to evaluate the optimal ponatinib dosing regimen for maximizing efficacy while mitigating toxicity. Results show that higher doses of ponatinib were associated with increased incidence of AOEs, with exposure-adjusted treatment-emergent AOE rates of 5.6%, 3.6%, and 2.1% for the 45-mg, 30-mg, and 15-mg cohorts, respectively [29]. However, the benefit differential was considerably larger with a starting dose of 45 mg, which was associated with a 26.3 percentage-point improvement in the response rate compared with a 15-mg starting dose (51.6% vs. 25.3%) [29]. Overall, the study indicated the best risk/benefit ratio when the 45-mg starting dose was reduced to 15 mg upon achievement of response (BCR::ABL1^IS transcript levels ≤ 1%) [29].

This retrospective study has strengths and limitations. The adjudication methodology provided a comprehensive and objective approach for characterizing AOE risk. A limitation is that only data from the clinical trial database were available. Prospective implementation of this strategy, as is being done in 2 ongoing trials, OPTIC and Ponatinib-3001 (NCT03589326), will overcome this challenge and add further value to the methodology and strength to the conclusions. In OPTIC, an independent cardiovascular endpoint adjudication committee is reviewing AOEs as they are reported using source documentation including cardiovascular workup (e.g., echocardiograms, electrocardiograms, biomarkers), hospitalization records, and any cardiovascular examinations performed.

Conclusions

Independent reconsideration of AOEs by an expert adjudication committee showed lower rates of clinically relevant AOEs overall (17% vs. 25%) and serious AOEs (16% vs. 20%) than were originally reported in the PACE trial, suggesting an earlier possible overestimation that may not accurately reflect the AOE risk with ponatinib. The incidence of exposure-adjusted newly occurring AOEs decreased over time during ponatinib treatment. Improved understanding of the AOE profile with ponatinib and risk factors for AOEs can help guide decisions around TKI treatment. Results from the OPTIC study support a novel ponatinib treatment regimen of a 45-mg starting dose reduced to 15 mg upon achievement of response, maximizing response while minimizing toxicity [29].