CADTH Reimbursement Review

Venetoclax (Venclexta)

Sponsor: AbbVie Corporation

Therapeutic area: Acute myeloid leukemia

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Introduction

Acute myeloid leukemia (AML) is a hematological malignancy defined by WHO as a myeloid neoplasm with greater than 20% blasts in the peripheral blood or bone marrow. Proliferating myeloid precursor cells leads to disruption of normal hematopoiesis and a clinical presentation of symptoms and complications of pancytopenia or leukostasis. Diagnosis is by complete blood count (CBC) and bone marrow biopsy, with identification of characteristic mutations and chromosomal rearrangements for targeted treatment and cytogenetic risk stratification. AML predominately occurs in older adults, with a median age of diagnosis of 67 years in Canada, and increasing incidence with age. Standard treatment for patients who are medically fit is intensive induction therapy with cytarabine and an anthracycline, but a substantial portion of patients with AML are ineligible for induction therapy due to frailty associated with age or comorbidities. Patients ineligible for treatment with induction therapy may be treated with hypomethylating agents (HMAs), such as azacitidine or low-dose cytarabine (LDAC), but rates of complete remission (CR) are low and duration of remission tends to be short. Prognosis for AML in older patients is poor, with 1 study reporting 5-year overall survival (OS) as 6.3% in patients aged 65 years.

Venetoclax is an orally administered highly selective inhibitor of the anti-apoptotic protein B-cell lymphoma 2 (BCL2). Health Canada granted a Notice of Compliance on December 4, 2020 for the following indication: Venclexta, in combination with azacitidine or LDAC, is indicated for the treatment of patients with newly diagnosed AML who are 75 years or older or who have comorbidities that preclude use of intensive induction chemotherapy. The recommended dose of venetoclax in combination with azacitidine is 400 mg/day for each day of a 28-day cycle following a 3-day ramp-up; azacitidine should be administered at 75 mg/m² for days 1 to 7 of the cycle. Dose adjustments are required in patients treated with strong and moderate inhibitors of CYP3A enzymes. Venetoclax has previously been reviewed by CADTH for its use in chronic lymphocytic leukemia (CLL) as monotherapy for patients with 17p deletion or without a 17p deletion who did not have other available treatment options in combination with obinutuzumab in previously untreated patients, and in combination with rituximab for patients who had received at least 1 prior therapy. A concurrent CADTH review of venetoclax with LDAC is ongoing.

The objective of the systematic review was to review the beneficial and harmful effects of venetoclax in combination with azacitidine for the treatment of patients with newly diagnosed AML who are 75 years or older, or who have comorbidities that preclude use of intensive induction therapy.

Stakeholder Perspectives

The information in this section is a summary of input provided by Canadian patient and clinician groups who responded to CADTH’s call for patient input and by clinical expert(s) consulted by CADTH for the purpose of this review.

Patient Input

One patient advocacy group, the Leukemia and Lymphoma Society of Canada (LLSC), provided input on venetoclax in combination with azacitidine for the treatment of AML. The LLSC used an online survey for its submission, which was conducted between December 7, 2020 and January 24, 2021. Twenty-nine patients responded, all from Canada, 5 of whom had experience with venetoclax in combination with azacitidine.

Many patients did not provide information on specific symptoms but described being diagnosed with AML as a life-changing event that affected not only themselves but their caregivers. Some patients needed to relocate to access treatment. Side effects of treatment, transfusion dependence and hospital admissions had a large impact on patients’ quality of life (QoL), as did isolation due to their vulnerability to infection. Patients reported the desired characteristics of treatment options as those that could maintain remission, were targeted with fewer side effects, covered by public plans, and accessible in their geographic region.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The experts indicted that currently available lower-intensity treatments have low rates of CR, and the CRs that are produced are not durable. They indicated that venetoclax plus azacitidine (or other HMAs) would change the current treatment paradigm, becoming the new standard of care for patients with treatment-naive AML who were ineligible for standard induction therapy, and providing an option for patients aged 75 years or older who were eligible for intensive chemotherapy, following discussion about risks and benefits.

The experts indicated that, at this time, there is insufficient information to make treatment decisions based on disease characteristics, and that while certain subgroups had been excluded from clinical trials, such as patients with central nervous system (CNS) involvement, these groups might reasonably be expected to benefit. The experts indicated that current evidence does not fully support the use of venetoclax plus azacitidine in fit patients eligible for standard induction treatment or in patients aged 75 years and older with good cytogenetic risk (core binding factor) AML who are fit for intensive induction chemotherapy, and their opinions differed in its suitability for patients with relapsed or refractory disease.

The experts indicated that response to treatment would be determined by achievement of CR with or without complete hematological recovery, as measured by CBC and bone marrow biopsy and/or transfusion independence or stable disease. OS and hospital visits, transfusion needs, and QoL were the most important end points. Assessment of response could be carried out after the first or second cycle.

The experts indicated that discontinuation of treatment might be determined by disease progression or intolerable adverse events, but could not comment whether venetoclax could be continued after azacitidine discontinuation. One respondent indicated that a bone marrow biopsy should be performed after the first and second treatment cycles, as response would be expected after a maximum of 2 cycles. Another indicated that response should be assessed at minimum after 4 to 6 cycles, but that most practitioners assess after the first cycle, given cost and to guide dosing of venetoclax for subsequent cycles.

The experts indicated that treatment should be given in a hospital or outpatient setting by a physician with experience looking after acute leukemia patients. Pharmacist involvement would be needed for management of drug interactions (e.g., azoles). Hospitalization might be required for ramping up the dose of venetoclax, with prophylaxis for tumour lysis syndrome, and the need for admission to manage neutropenic fever and other complications during therapy should be anticipated.

Clinician Group Input

Four clinician groups provided input: the Canadian Leukemia Study Group (CLSG), the Ontario Health (Cancer Care Ontario) Hematology Disease Site Drug Advisory Committee (OH-CCO Hem-DAC), the Leukemia/Bone Marrow Transplant (L/BMT) Program of British Columbia, and the Alberta Tumour Board Myeloid Physicians Group (ATB-MPG).

There were no substantive differences in opinions between the clinical experts consulted by CADTH and the clinical groups. The groups noted that patients are aware of venetoclax and azacitidine, and some patients have been “self-funding” venetoclax by using CYP3A inhibitors to reduce the dose and, thus, the cost of venetoclax.

Drug Program Input

The drug programs indicated that current treatment options for patients with newly diagnosed AML who are ineligible for intensive chemotherapy include azacitidine, LDAC, and best supportive care (BSC). The reimbursement of venetoclax plus azacitidine would likely replace azacitidine in this treatment setting. Azacitidine is funded in most jurisdictions for patients with AML who are ineligible for intensive chemotherapy, and some jurisdictions fund alternate dosing schedules for azacitidine (i.e., 5 to 2-2 and 6 consecutive days) in addition to the schedule of 7 consecutive days. However, it was noted that some patients 75 years of age and older may be fit to tolerate intensive chemotherapy. The ramp-up dosing schedule for venetoclax with azacitidine differs significantly from the ramp-up dosing schedule already in use for chronic lymphocytic leukemia (CLL) indications and the current packaging for venetoclax is designed for the CLL ramp-up dosing schedule. Venetoclax plus azacitidine includes an oral and an IV and subcutaneous drug and therefore would be reimbursed through different programs in some jurisdictions. The drug programs identified the potential for indication creep for patients with a high risk of myelodysplastic syndrome (MDS), those who have progressed or have had an inadequate response on low-dose chemotherapy for AML, and patients who have relapsed after induction chemotherapy and are not eligible for stem cell transplant and who are then treated with azacitidine. It was noted that treatment combination increases the need for health care resources (i.e., hospital admission and additional pharmacy and nursing resources for the potential management of tumour lysis syndrome and monitoring for drug interactions). Affordability was also identified as an issue since the combination is expected to replace azacitidine monotherapy.

Clinical experts were consulted by CADTH for questions related to implementing venetoclax plus azacitidine into current provincial drug plans. Overall, most implementation questions related to the dosing schedule and administration and the eligible patient population.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of Studies

One double-blind, placebo-controlled phase III randomized controlled trial (VIALE-A) contributed evidence to this review. The trial objective was to evaluate the efficacy and safety of venetoclax plus azacitidine compared with placebo plus azacitidine in adults with newly diagnosed AML who were 18 years or older and ineligible for standard induction therapy due to age or comorbidities. The trial was restricted to patients who had not previously been treated with an HMA and who had intermediate or poor risk cytogenetics. The primary outcomes were OS and composite complete remission rate (i.e., CR plus complete remission with incomplete marrow recovery [CR + CRi]). Secondary outcomes were CR, CR plus complete remission with incomplete hematological recovery [CR + CRh], rate of CR + CRi by the initiation of cycle 2, transfusion-independence rate, minimal/measurable disease (MRD) response rate, response rates and OS in molecular subgroups, fatigue, global health status and quality of life (GHS/QoL), and event-free survival (EFS).

A total of 431 patients were randomized in a 2:1 ratio: 286 to venetoclax plus azacitidine and 144 to placebo plus azacitidine. The most common reasons given for patients to be considered ineligible for standard induction therapy were age and Eastern Cooperative Oncology Group Performance Status (ECOG PS). Patients were elderly, with poor performance and markers of severe disease. The mean age was 75.4 years, with 60.6% aged 75 years or older. Almost all patients were White or Asian, and the majority of patients were male (60.1%). Most (75.2%) had de novo rather than secondary AML. Nearly 2-thirds had intermediate risk cytogenetics, 1-third had poor risk, and 1-half had bone marrow blasts of 50% or greater at baseline.

Efficacy Results

Table 2 shows a summary of the key efficacy and safety outcomes. Venetoclax plus azacitidine improved most outcome measures that were identified as being of interest to clinicians and patients. Statistically significant treatment differences were seen for OS, EFS, measures of disease response (CR + CRi, CR + CRh, CR), and post-baseline transfusion independence. Improvements were also seen for OS and CR + CRi in the subgroup of patients with isocitrate dehydrogenase 1 (IDH1) or IDH2 mutation and for CR + CRi among patients with FMS-like tyrosine kinase 3 (FLT3) mutations. No statistically significant difference was detected in OS for patients with FLT3 mutations; however, the subgroup was small, making it difficult to detect a difference. While clinically meaningful differences in patient-reported outcomes of GHS/QoL and fatigue were observed at individual end points, differences between treatment groups could not be interpreted because the sequential testing strategy failed before this level.

Harms Results

Table 2 shows a summary of the key efficacy and safety outcomes. All patients in both groups experienced at least 1 adverse event, and almost all experienced at least 1 grade 3 or greater adverse event. Compared with patients who received placebo plus azacitidine, a greater proportion of patients who received venetoclax plus azacitidine experienced 1 or more serious adverse events (SAEs), 1 or more adverse events leading to discontinuation or dose interruption of venetoclax or placebo or azacitidine, or 1 or more adverse events leading to death. Common harms in all categories are generally predictable from the known mechanism of action for venetoclax and/or azacitidine and the underlying disease. Cytopenias were common, with neutropenia, febrile neutropenia, thrombocytopenia, and anemia represented across all categories, as were gastrointestinal adverse effects. Febrile neutropenia and infections contributed substantially to most common SAEs and were the most frequent adverse events leading to death.

The notable harms identified for the protocol were neutropenia, febrile neutropenia, infections, tumour lysis syndrome, hemorrhage, and secondary malignancies. Neutropenia, febrile neutropenia, infections and infestations, and secondary primary malignancies all occurred in a greater proportion of patients who received venetoclax plus azacitidine than in patients who received placebo plus azacitidine. Hemorrhage and tumour lysis syndrome occurred in similar proportions, and the proportion of patients with tumour lysis syndrome was low (≤ 2.5%). The most common secondary malignancies were basal cell carcinoma and squamous cell carcinoma of the skin.

Critical Appraisal

The study was well conducted, with no clinically meaningful imbalance in baseline characteristics, minimal loss to follow-up, and a collection of end points that were standardized and meaningful to patients. Multiplicity was controlled throughout testing of the primary and secondary efficacy end points, with pre-specified strategies for testing of end points. The overall rate of discontinuations from the study was low and assumptions surrounding missing data were conservative for most end points. Interpretation of patient-reported outcome data is limited due to attrition of numbers over cycles.

The generalizability concerns that were identified included the assumption that patients aged 75 years and older would not be eligible for standard induction therapy, and the need for venetoclax and azacitidine to be limited to settings that could provide monitoring and supportive care. In the Canadian setting, patients aged 75 years and older would be considered for treatment if they were medically fit, especially if they had good or intermediate risk cytogenetics. Patients from rural and remote Canadian settings would have to travel for care or would be limited to other treatment options.

Indirect Comparisons

Description of Studies

A systematic review was conducted of trials comparing venetoclax plus azacitidine, venetoclax plus LDAC, azacitidine alone, LDAC alone, and BSC in adults with AML who were not eligible for standard induction chemotherapy. Three indirect treatment comparison (ITC) analyses were conducted: 1 network meta-analysis (NMA) and 2 propensity score–weighting analyses that compared venetoclax plus azacitidine with LDAC and azacitidine with LDAC. For the NMA, HR data were available for OS for 4 trials in a connected network and for proportions of patients with CR + CRi for 3 trials. For the propensity score–weighting analysis, data were available for OS, EFS, and CR + CRi from VIALE-A and the LDAC group from VIALE-C.

Efficacy Results

In the NMA, the results favoured a lower hazard of death for patients assigned to venetoclax plus azacitidine compared with azacitidine (HR = 0.66; 95% credible interval [CrI], 0.52 to 0.85), LDAC (HR = 0.57; 95% CrI, 0.40 to 0.81), and BSC (HR = 0.37; 95% CrI, 0.24 to 0.58), with no treatment favoured between venetoclax plus azacitidine and venetoclax plus LDAC (HR = 0.81; 95% CrI, 0.50 to 1.31). For CR + CRi, venetoclax plus azacitidine was favoured over azacitidine (odds ratio [OR] = 5.05; 95% CrI, 3.30 to 7.87), LDAC (OR = 5.42; 95% CrI, 2.80 to 10.50), and BSC (OR = 61.55; 95% CrI, 8.23 to 1,881.53), with no treatment favoured between venetoclax plus azacitidine and venetoclax plus LDAC (OR = 0.86; 95% CrI, 0.30 to 2.35).

In the first propensity-score analysis, venetoclax plus azacitidine was favoured over LDAC for OS (HR = 0.50; 95% confidence interval [CI], 0.35 to 0.73), EFS (HR = 0.40; 95% CI, 0.28 to 0.58), and CR + CRi (OR = 10.17; 95% CI, 4.55 to 22.73). In the second propensity-score analysis for OS, venetoclax plus azacitidine was favoured over LDAC (HR = 0.52; 95% CrI, 0.36 to 0.77) and azacitidine (HR = 0.64; 95% CrI, 0.50 to 0.82), and no statistically significant difference was seen between azacitidine and LDAC (HR = 0.78; 95% CrI, 0.52 to 1.17). For EFS, venetoclax plus azacitidine was favoured over azacitidine (HR = 0.62; 95% CrI, 0.49 to 0.77) and LDAC (HR = 0.41; 95% CrI, 0.29 to 0.59), and azacitidine was favoured over LDAC (HR = 0.63; 95% CrI, 0.43 to 0.92). For CR + CRi, venetoclax plus azacitidine was favoured over azacitidine (OR = 5.02; 95% CrI, 3.24 to 7.77) and LDAC (OR = 9.69; 95% CrI, 4.30 to 21.85), and no statistically significant difference was seen between azacitidine and LDAC (OR = 1.93; 95% CrI, 0.82 to 4.54).

Harms Results

No analysis of harms was included in the indirect comparisons.

Critical Appraisal

A key limitation of the NMA was the clinical heterogeneity between studies in potential treatment-effect modifiers of blast count at baseline, prior treatment with an HMA, and cytogenetic risk. As the network was sparse, fixed-effects models had to be used, and there was no opportunity for baseline covariate adjustments. Due to these limitations, the comparative efficacy estimates may be biased, and it is not possible to quantify or identify the direction of the bias. Certain estimates, particularly for CR + CRi, were imprecise due to sparse data. In the propensity-score analyses, weighting was generally good, but the relatively small numbers of patients in the LDAC comparator group limited the number of covariates that could be included in the model. The comparisons were not randomized and the results were highly susceptible to bias due to imbalances in unmeasured confounders.

Conclusions

One double-blind, placebo-controlled phase III randomized clinical trial (RCT) (VIALE-A) and 1 ITC provided evidence supporting the efficacy and safety of venetoclax plus azacitidine in adult patients ineligible for standard induction chemotherapy due to age or comorbidities. Compared with azacitidine alone, patients treated with venetoclax (400 mg daily) and azacitidine (75 mg/m² on days 1 through 7 of a 28-day cycle) showed benefits in important clinical end points of OS, overall and early composite complete remission (CR + CRi), EFS, CR, and transfusion-independence data (red blood cell or platelet). All study participants reported treatment-emergent adverse events. For most categories of adverse events, there was an overall higher proportion of patients reporting in venetoclax plus azacitidine. The most common adverse events were cytopenias and infections. No firm conclusions can be drawn for differences between groups in GHS/QoL and fatigue, and patient attrition reduced the number of observations over the cycles, which limits the interpretation for these end points. Overall, the study was well conducted.

The VIALE-A study did not include a comparison between venetoclax plus azacitidine and current standards of care of induction therapy (in patients aged ≥ 75 and fit), LDAC, or BSC, or the alternative combination of venetoclax plus LDAC. In an ITC, venetoclax plus azacitidine was favoured over monotherapies and basic supportive care, but no treatment was favoured for survival or for composite complete remission between venetoclax plus azacitidine and venetoclax plus LDAC. No data are available for the comparison of venetoclax plus azacitidine with induction therapy. Results for 2 propensity-score comparisons between venetoclax plus azacitidine and azacitidine and LDAC were consistent. Small study and patient numbers and the potential for bias limit the reliability of the ITC, and the propensity-score comparisons were not randomized and therefore highly susceptible to bias.

Introduction

Disease Background

AML is a hematological malignancy defined by WHO as a myeloid neoplasm with greater than 20% blasts in the peripheral blood or bone marrow.² AML results from malignant transformation of myeloid precursor cells to produce 1 or more clonal populations that can proliferate but do not normally differentiate into their mature forms. This leads to an accumulation of leukemic blasts or immature cells in the bone marrow, peripheral blood, and extramedullary tissues, which disrupt normal hematopoiesis. WHO 2016 guidelines² identify 6 distinct groups of AML:

• AML with recurrent genetic abnormalities

• AML with myelodysplasia-related changes

• therapy-related myeloid neoplasms

• AML not otherwise specified

• myeloid sarcoma

• myeloid proliferations related to Down syndrome

AML is the most common form of acute leukemia in adults. According to the Canadian Cancer Society’s most recent data, in 2016, 1,090 people were diagnosed with AML, 610 men and 480 women.³ In 2017, 1,184 people died of AML, 678 men and 506 women.³ Projections for the Canadian population in 2020 only report figures for all forms of leukemia; according to these projections, 6,900 patients would be diagnosed with leukemia, and 3,000 would die. Assuming that around 24% of leukemia in Canada is AML (data from 1992 to 2008),⁴ the subgroup diagnosed with AML would comprise approximately 1,660 patients.

AML occurs predominately in older adults, with a median age at diagnosis of 67 years in Canada.⁴ Incidence increases with age. There is an increased incidence in men and non-Hispanic Whites compared with women and other racial or ethnic groups. Certain environmental exposures have been associated with an increased risk of AML (e.g., chemicals, radiation, tobacco, and retroviruses). AML can occur as a secondary malignancy following chemotherapy, or develop out of a pre-existing hematopoietic abnormality such as MDS, other myeloproliferative neoplasms, paroxysmal nocturnal hemoglobinuria, aplastic anemia, or clonal hematopoiesis of indeterminate prognosis. Rarely, it can be associated with an inherited genetic abnormality or familial predisposition to hematologic disorders.

Patients typically present with the symptoms or complications of disrupted hematopoiesis (bleeding or bruising, infection that can be life-threatening, fatigue or shortness of breath, headache or focal neurologic complaints) or symptoms of leukostasis resulting from an excess of immature white cells in the peripheral blood. A presumptive diagnosis of leukemia may be made from a CBC and smear, but confirmation is usually by bone marrow biopsy and aspirate. Characterization of specific genetic abnormalities enables stratification by genetic risk into favourable, intermediate, and poor risk categories, which is used to guide treatment decisions or use of therapies targeted against specific mutations.

The overall 5-year survival for AML in adults in Canada is 21%.⁵ Older patients have notably poorer survival. In US data, the 5-year survival is 44.8% in patients younger than 65 years and 6.3% in patients aged 65 years and older.⁶ In another study, patients aged 60 years and older had 1-year and 5-year survival of 20.1% and 8.4%, respectively.⁷ Survival is influenced by cytogenetic and genetic risk.

Standards of Therapy

Standard treatment for patients who are medically fit consists of cytotoxic remission induction therapy with cytarabine, administered by infusion over 7 days, combined with an anthracycline, usually daunorubicin or idarubicin, given daily for the first 3 days. Induction therapy is followed by high-intensity consolidation therapy. This may be accompanied by targeted therapy for specific clinical situations or genetic mutations: midostaurin in patients with FLT3, and gemtuzumab ozogamicin (monoclonal antibody against CD33) in patients with favourable and intermediate risk disease.⁸

Determination of eligibility for intensive chemotherapy is based on patient age, fitness, presence of comorbidities, and patient preferences. In general, intensive therapy is poorly tolerated by older patients. According to the 2017 Canadian Consensus Guidelines for treatment of older patients with AML⁸ induction therapy shows a survival benefit for patients up to age 80, with the exception of those with major comorbidities or those with adverse risk cytogenetics who were not candidates for hematopoietic stem cell transplant.⁸ Anthracycline and cytarabine are the recommended drugs for induction therapy, with the addition of midostaurin for patients with an FLT3 mutation, and the addition of gemtuzumab ozogamicin for patients with de novo AML and favourable or intermediate risk cytogenetics. For patients who are not eligible for induction therapy, azacitidine is recommended for those with adverse risk cytogenetics or transformed from MDS, while either HMA or LDAC could be used for others. Acute promyelocytic leukemia would be treated with arsenic trioxide plus all-trans retinoic acid (with an anthracycline for those with white blood cell count > 10 × 10⁹/L).

Azacitidine has been approved by Health Canada for patients with low blast count AML (blast counts of 20% to 30%); however, in multiple jurisdictions, it is used and provides clinical benefit in patients with blast counts of 30% or greater. Some jurisdictions fund alternative dosing schedules for azacitidine besides the standard 7-day consecutive regimen. According to input from the clinicians consulted by CADTH for the purpose of this review, in real-world clinical practice in Canada, many eligible patients were unable to receive azacitidine-based therapy, as this drug has to be administered in an oncology clinic setting because of its instability after reconstitution. Many patients who live in rural areas and some in urban settings are unable to travel to these clinics regularly to receive treatment due to the distances involved, their overall frailty, and challenges in obtaining suitable transportation.⁹

Not all patients respond to first-line therapy and all become refractory to current treatment options, with limited life expectancy. There are few effective treatment options following relapse on front-line AML therapy⁸; some patients will receive off-label azacitidine with or without venetoclax or participate in a clinical trial. A minority of patients with FLT3 mutations receive gilteritinib, but many patients are not well enough to tolerate further therapy; hence, they receive BSC only.⁸

Drug

Venetoclax is an orally administered highly selective inhibitor of the anti-apoptotic protein BCL2. Increased expression of BCL2 has been measured in AML blasts, and a majority of AML stem cells express high levels of BCL2 and are dependent on BCL2 for survival. High levels of expression of BCL-2 have been associated with poorer response to chemotherapy and poorer survival in patients with AML. Azacitidine is also thought to affect the inhibition of the pro-survival proteins MCL1 and BCL-XL, so co-administration of azacitidine should increase the dependence of leukemia cells on the BCL2 pathway for survival, and potentiate the effect of venetoclax.

Venetoclax was granted a Health Canada Notice of Compliance on December 4, 2020. The approved indication was: Venclexta (venetoclax) in combination with azacitidine or LDAC, is indicated for the treatment of patients with newly diagnosed AML who are 75 years or older, or who have comorbidities that preclude use of intensive induction therapy. This is consistent with the reimbursement request. Venetoclax has also been approved for the treatment of CLL, as monotherapy for patients with or without a 17p deletion who do not have other available treatment options, in combination with obinutuzumab in previously untreated patients, and in combination with rituximab for patients who have received at least 1 prior therapy. A concurrent CADTH review of venetoclax with LDAC is ongoing, and previous CADTH reviews were conducted for the indications in CLL.

Table 3 summarizes the characteristics and indications for venetoclax, azacitidine, and cytarabine.

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

About the Patient Group(s) and Information Gathered

One patient advocacy group, the LLSC, provided input on venetoclax in combination with azacitidine for the treatment of AML.

The LLSC’s mission is to cure leukemia, lymphoma, Hodgkin’s disease, and myeloma, as well as to improve the QoL of all Canadians affected by blood cancers. The LLSC has received funding from AbbVie.

The LLSC used an online survey for its submission, which was conducted between December 7, 2020 and January 24, 2021.

A total of 29 patients responded. All respondents were from Canada: 13 from Ontario, 6 from Quebec, 6 from British Columbia, and 4 from Alberta. Patient ages ranged from 25 to 84 years old and 2 were 75 years old or older. There were 18 females and 10 males and 1 did not report gender. Comorbidities were not reported. All patients had been diagnosed with AML within the past 7 years. Five of the respondents had experience with venetoclax in combination with azacitidine.

Disease Experience

According to the patient respondents to the LLSC survey, the symptoms that patients with AML experience that impact QoL include fatigue, suddenness of symptom development, anxiety, fear of relapse (number of patients unspecified for preceding symptoms), and loss of eyesight (n = 1). One patient experienced a spleen rupture and was in a coma for 8 days. Fatigue was the symptom mentioned most often among patient respondents. Fatigue and other symptoms subsequently affected social and family life along with other symptoms. Patients reported that these symptoms compounded with the changes related to the coronavirus disease 2019 (COVID-19) pandemic, which led to further social isolation. Some patients reported they are unable to work due to their disease and associated symptoms. Many patients did not provide information on the specific symptoms they experienced but described being diagnosed with AML as a life-changing event. Below are comments from patients regarding their experiences with AML:

“Everything in my life stopped cold turkey-employment, social life, relationships, etc. I made a complete personal 360 degree pivot to focus on my healing and living.”

“Well COVID and my compromised immune system has caused me to be very socially isolated. I haven't seen some very important people in my life for almost 2 years at this point.”

When asked if there are any aspects or symptoms of AML that are easier to control, most patients (n = 7) indicated no, and 1 patient commented there was no control with AML. Three patients indicated exercise was helpful in alleviating some symptoms, reporting that exercise and keeping physically active helped, particularly with fatigue.

Two patients reported feeling no impact or back to normal at the time of survey.

AML affects not just those who are diagnosed, but also their caregivers, which may include a spouse, immediate family members, and friends. Patients reported needing assistance for physician visits and daily activities. According to the LLSC survey, patients reported that caregivers seemed to feel multiple emotions about the patient’s AML: stress, worry, sadness, insecurity, and fear of dying were all frequently mentioned. Their companion through the disease journey was important for patients.

Experiences With Currently Available Treatments

According to the LLSC survey, the front-line treatments that patients received after diagnosis included chemotherapy (n = 24), stem cell transplant or bone marrow transplant (n = 16), drug therapy (n = 6), radiation therapy (n = 5), and chimeric antigen receptor (CAR) T-cell therapy (n = 1). One patient reported receiving Vyxeos (daunorubicin and cytarabine). Patients reported a wide range of side effects with current treatments, and the ones they considered to have a large impact on their QoL included hair loss (n = 17); weakness (n = 15); extreme fatigue (n = 14); diarrhea (n = 10); infections (n = 8); anemia (n = 8); mouth sores (n = 8); nausea and vomiting (n = 7); fever (n = 6); low blood cell counts (n = 6); tingling sensations (n = 4); constipation (n = 2); graft-versus-host disease (n = 2); lung, heart, kidney, or nerve problems (n = 2); cough (n = 1); rashes (n = 1); shortness of breath (n = 1); and psychological distress (n = 1). The side effects due to chemotherapy and stem cell transplant had a large impact on patients’ QoL, summarized the LLSC survey administrator. These side effects from front-line treatments led to changes in physical activity (n = 15), anxiety (n = 11), problems in mental health and overall happiness (n = 11), eating challenges (n = 12), and social development (n = 6) and educational development (n = 6) challenges. Overall, the side effects from front-line treatments caused significant disturbance to daily living. Patients were isolated from visitors during stem cell transplant. Opportunistic infection could occur. The following are comments from patients regarding their experiences with front-line AML treatments:

“The main challenge was the nausea and vomiting. I didn’t seem to have much control over it and had my wonderful bucket always with me. I could be fast asleep and awake and vomit.”

“Your whole world changes when you are diagnosed with AML. Suddenly, you confront your mortality. You feel extremely weak, you have to go into hospital for months, and you don't realize you MUST go into remission to have a stem cell transplant.”

“Extremely tired and little desire to be active. Difficulty eating and keeping it down. A few days of low hemoglobin and fluid on the lung that caused shortness of breath.”

“The worst issue is that I have no more job and that the treatments made me lose a lot of concentration and I get exhausted easily.”

“Had to move to Vancouver for treatment for 9 months. 2 or 3 months total in hospitals. Daily outpatient care. Kinda turns your life upside-down.”

Patients who responded to the LLSC survey reported a mixture of both positive and negative experiences accessing treatments. Thirteen respondents reported generally positive experiences and some patients attributed their experience to the support from medical staff. Six patients reported negative experiences. Negative experiences were related to challenges with receiving care and treatment plans. Some patients needed to relocate to receive treatments.

Improved Outcomes

The majority of respondents to the LLSC survey indicated that the factors they considered about a new cancer treatment were physician recommendation (n = 19), possible impact on disease (n = 17), QoL (n = 12), closeness of home to the treatment centre (n = 9), and outpatient treatment (n = 8).

The LLSC survey patient respondents also reported the characteristics of new treatment options they hoped to have, particularly those that could maintain remission, have fewer side effects, be covered by public plans, and be accessible in their geographic region. The opportunity to have access to other supportive options, such as meditation, hypnosis, neuro-linguistic programming support, and awareness support (thoughts, emotions, and behaviours), was also mentioned.

Experience With Drug Under Review

Five of the LLSC respondents indicated they took venetoclax in combination with azacitidine. Three respondents received it through compassionate use from the pharmaceutical company and 2 respondents received it through physician prescription. Two respondents reported difficulties in accessing the drug because of the costs and 1 reported a long wait time. The financial costs reported by 3 patients ranged from $10,000 to $13,000. Three respondents took this treatment because of physicians’ recommendation and 2 patients considered side effects to be the main factor for adopting this treatment. One of them also considered the chance of survival the reason to take this treatment.

Patients’ responses to this treatment varied greatly, including an overall great experience (n = 1), experiences with side effects (tiredness and loss of appetite, n = 1), no side effects but relapse (n = 1), significant side effects (n = 1), and transition to transplant (n = 1). Serious and very serious side effects reported by patients were fatigue (n = 3), low platelets (n = 2), and anemia (n = 1). Other side effects were minor or manageable, such as diarrhea, nausea, constipation, cough, back pain, and headache. One patient thought multiple visits to the hospital, transfusion, and infection made managing the side effects more challenging.

Two patients strongly agreed that venetoclax in combination with azacitidine improved their QoL compared with other treatments they received. Two patients also agreed and 1 disagreed. Two patients thought this treatment led to remission and thus improved their QoL. Overall, patients’ experiences with this treatment varied greatly. Compared with other treatments they received, patients indicated this treatment was significantly more challenging (n = 2), more challenging (n = 1), neutral (n = 1), less challenging (n = 1), and significantly less challenging (n = 1). The 5 patients were willing to tolerate the side effects of this treatment. One patient tolerated the side effects to live and another thought there was no other choice.

Clinician Input

All CADTH review teams include at least 1 clinical specialist with expertise regarding the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 3 clinical specialists with expertise in the diagnosis and management of AML in adults.

Unmet Needs

The experts indicated that current lower-intensity treatments have low rates of CR, and CRs that are produced are not durable. Treatments producing higher rates of CR had increased toxicity and are not tolerable in the patient population under study. The Health Canada approval for azacitidine is only for treatment-naive patients with higher-risk MDS and AML (according to the International Prognostic Scoring System) and with up to 30% blasts by WHO classification, although certain local jurisdictions make azacitidine available for patients with 30% or greater blasts.

Place in Therapy

The experts indicated that venetoclax plus azacitidine (or other HMA) would change the current treatment paradigm. It would be the new standard of care for patients with treatment-naive AML aged 18 years or older who are not eligible for intensive chemotherapy, and would replace, for the most part, single-agent HMAs or LDAC. For patients with treatment-naive AML aged 75 years or older who were eligible for intensive chemotherapy, especially those with good or intermediate risk cytogenetics, there would have to be a discussion with the patient about the risks and benefits of the different treatment options. It should be noted there is no consistency as to the upper age limit at which an acute leukemia treatment centre would administer intensive chemotherapy. As venetoclax plus azacitidine is myelosuppressive, it may not be suitable for a small number of frail patients, or for those who would be unable to travel to the treating hospital for count checks. This, too, would need to be assessed by the treating physician in conjunction with the patient.

Patient Population

The experts indicated that patients must have a diagnosis of AML with greater than 20% blasts, and that making this diagnosis from blood and bone marrow is straightforward. Patients with isolated granulocytic sarcoma were not included in the trial.

The clinicians indicated that selection of patients for treatment would be based on clinician judgment and patient preferences. At this time, there is not enough information to make treatment decisions based on disease characteristics, and the data to predict response have not yet been validated in large studies.

Patients with good risk cytogenetics and patients with myeloproliferative neoplasm in blast crisis have been excluded from studies of venetoclax plus azacitidine or LDAC. Patients who had previously used an HMA were not eligible for the VIALE-A trial, but were eligible for the VIALE-C trial. One respondent indicated that studies suggest response to venetoclax plus HMA following HMA is similar to response to venetoclax plus LDAC; however, there are no direct comparisons of these 2 treatments post HMA. Patients with CNS involvement by AML have been excluded from all AML studies, but this does not mean this group of patients would not benefit from venetoclax plus azacitidine with concomitant intrathecal therapy, similar to the current practice of administering systemic intensive chemotherapy and intrathecal therapy to those patients who have CNS involvement by AML.

One expert suggested venetoclax with azacitidine was preferred over venetoclax plus LDAC in patients who had not received prior HMA. Venetoclax plus HMA was reasonable in patients with prior HMA use, and ivosidenib plus azacitidine would be reasonable in patients with IDH1 mutations, if ivosidenib were available.

The experts indicated that venetoclax plus azacitidine would not be suitable for fit patients (i.e., those eligible for standard induction treatment) with good cytogenetic risk AML, or for patients with acute promyelocytic leukemia. Emerging evidence may support the use of venetoclax plus azacitidine in fit patients as a bridge to allogeneic bone marrow transplant in higher-risk AML. Opinions differed in the use of venetoclax plus azacitidine in patients with relapsed or refractory disease, as venetoclax plus azacitidine has reduced activity in relapsed and refractory disease: 1 respondent did not recommend it, and another thought it was a reasonable option.

Assessing Response to Treatment

The experts indicated that, in clinical practice, response to treatment would be determined by achievement of CR, CRh, or CRi and/or transfusion independence or hematological improvement or stable disease. A clinically meaningful response to treatment would be represented by improved OS, improved EFS, achievement of durable CR, decreased hospitalizations, decreased transfusion requirements, slower progression, stabilization of disease (which would presumably improve or not worsen symptoms), and improved QoL. One clinician noted that the strict definitions of response did not necessarily identify responding patients. One clinician indicated that OS, hospital visits, transfusion dependence, and QoL were likely the most important end points. The clinicians indicated it is difficult to determine a minimum improvement over standard of care and what a meaningful response represents will vary by physician and patient.

Response would be measured by CBC and bone marrow blasts. One respondent indicated bone marrow biopsy should be performed after the first and second cycles of treatment, as response would be expected after a maximum of 2 cycles. Another indicated that response should be assessed at minimum after 4 to 6 cycles, but that most practitioners assess after the first cycle, given cost and to guide dosing of venetoclax for subsequent cycles. Once a response was obtained, then CBC could be followed for evidence of progression.

Discontinuing Treatment

The experts agreed that disease progression and intolerable adverse events were factors to be considered in the decision to discontinue treatment. Disease progression could be indicated by worsening CBC, increasing blasts in bone marrow, and/or loss of transfusion independence. The clinicians could not comment whether venetoclax could be continued if a patient stopped azacitidine.

Prescribing Conditions

The experts indicated that a hospital or outpatient clinic would be appropriate settings for treatment. As venetoclax plus azacitidine is myelosuppressive, physicians should have experience in looking after acute leukemia patients. Patients might require hospital admission for venetoclax dose ramping up. The proportion depends upon population: 1 respondent indicated the proportion would small, and another that it could be 25% to 50%. Patients would also require pre-treatment and monitoring for tumour lysis syndrome (occurring in 1% to 2% of patients). A not-insignificant proportion of patients will need to be hospitalized for neutropenic fever and other complications during their cycle of therapy. Pharmacists would be involved in reviewing medications, as a significant proportion of patients are on azoles, which interact with venetoclax and require dose modifications.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups.

Four clinician groups provided input on the reimbursement review of venetoclax in combination with an HMA or in combination with LDAC, for the treatment of adult patients with newly diagnosed AML who are ineligible for intensive chemotherapy.

The ATB-MPG is a group of physicians who treat myeloid malignancies and acute leukemias (myelodysplastic syndromes, myeloproliferative neoplasms, AML, and acute lymphoblastic leukemia) within Alberta and function as a group within the Alberta Hematology Tumour Group. There are Edmonton and Calgary groups that meet regularly and provincially every 3 months as well as annually to update the treatment guidelines on these diseases for Alberta. Information was collected by ATB-MPG for this review through literature review and group discussions. The group reviews written guidelines in a group setting and modifies its opinions based on written and oral discussion among the members. The discussed information is then approved by the full group.

The CLSG is a cross-Canada collective of physicians who treat acute leukemia and who represent all major leukemia centres in all provinces. The purpose of CLSG is “to improve the diagnosis and treatment of leukemia in Canada by identifying diagnostic and management best practices, promoting Canada-wide standards of care, fostering clinical and basic leukemia research, and improving new drug access.” Information for this review was gathered through ongoing group discussions and polling of members, with input requested from other international experts as appropriate. The written opinions were further edited and approved by the full group.

Ontario Health’s (Cancer Care Ontario) Hematology Disease Site Drug Advisory Committee (OH-CCO Hem-DAC) provides evidence-based clinical and health system guidance on drug-related issues in support of OH-CCO’s mandate, including provincial drug reimbursement programs and the Systemic Treatment Program. The group gathered its information for this review through discussions at the DAC meeting.

The L/BMT Program of British Columbia is a joint program of BC Cancer and Vancouver Coastal Health with a primary mandate for the province to treat acute leukemia, perform stem cell transplantation, and deliver cellular therapies for patients with hematologic malignancies. Members from the acute working group within the program provided input for this review. The group primarily reviewed published data from the phase III trial (DiNardo)²³ as well the data from the phase Ib trial and reviewed National Comprehensive Cancer Network (NCCN) guidelines.

Unmet Needs

All 4 clinician groups agreed that the current treatment paradigm for this disease includes azacitidine, LDAC, and BSC.

In addition, the L/BMT program of British Columbia noted that determination of eligibility for intensive chemotherapy is based on patient age, fitness, presence of comorbidities, and patient preferences. The group added that in British Columbia, azacitidine is given at many BC Cancer sites as well as in other hospital outpatient settings depending on the geographic location and treating physician. The treatment is given as a subcutaneous injection usually in either a 7-day or 5 days plus 2 days schedule. L/BMT members added that response to azacitidine is not evident before cycles 3 to 4 and is usually formally assessed around cycles 6 to 7 with a repeat bone marrow aspirate and biopsy. The group added that treatment with azacitidine continues indefinitely as long as the patient benefits and tolerates the treatment. L/BMT noted that LDAC is also given as a subcutaneous injection of 20 mg twice a day for 10 days every 4 to 6 weeks and can be given by the patient or a caregiver at home and also requires patient education by a chemotherapy-trained nurse before initiation. The group noted that LDAC is given less frequently than azacitidine and is beneficial for patients who live a long distance from a cancer centre that administers azacitidine or who would prefer to receive treatment at home. L/BMT also noted that in the province of British Columbia, LDAC is generally reserved for patients with intermediate risk karyotype. L/BMT added that older patients with AML also receive supportive treatments either with azacitidine or LDAC or alone, and these supportive treatments include transfusion support, hydroxyurea, antibiotic treatment, pain control, and palliative care. The group noted that currently there are no relevant special access programs available and that novel treatments for AML and non-intensive treatments with LDAC and azacitidine can improve symptoms and result in clinical responses, with a small proportion of patients achieving CR. L/BMT added that this treatment (venetoclax plus azacitidine) is associated with an improvement in OS, but it is not considered curative and responding patients will ultimately have disease progression.

The ATB-MPG added that, currently, for AML patients who are ineligible for induction chemotherapy, the common clinical practice for the majority of patients is to use azacitidine 75 mg/m² per day for 7 days every 28 days, and some patients receive cytarabine 20 mg twice daily for 10 days. For elderly patients, the group noted there was no survival benefit noted with cytarabine, and azacitidine is preferred. They also added that azacitidine is approved in Canada for patients with low blast count AML (20% to 30%), but is commonly used and provides clinical benefit to patients unfit for induction chemotherapy who have blast counts higher than 30%. In addition, the group noted there is a temporary compassionate access to an oral decitabine and cedazuridine compound that is available to patients with low blast count AML (20% to 30%); however, ongoing access to this drug is not yet established. The ATB-MPG agreed that palliative basic supportive care options include hydroxyurea and blood transfusion support as well as antibiotics, and patients may be offered clinical trials when they are available. The group noted that patients are aware of venetoclax and azacitidine and some have been “self-funding” venetoclax by using CYP3A inhibitors to reduce the dose, and thereby the cost, of venetoclax.

CLSG agreed with the other clinician groups that approximately 40% to 50% of newly diagnosed AML patients are judged to be unfit for intensive induction chemotherapy and this includes patients aged 75 and older as well as younger patients with severe comorbidities. For these patients, CLSG noted that the treatment options include azacitidine, LDAC, or BSC alone. The group added that for patients with poor risk cytogenetics or AML transformed from MDS, azacitidine is the current treatment of choice, while for patients with AML arising de novo with standard risk cytogenetics, azacitidine or LDAC can be used. The group noted that in the real-world clinical setting, many patients in Canada are not able to receive azacitidine-based therapy, as this drug needs to be given in an oncology clinic setting due to its instability after reconstitution. As a result, many patients who live in rural areas and some in urban settings are unable to travel to these clinics regularly to receive treatment because of the distances, overall frailty, and challenges in obtaining suitable transportation.

All 4 clinician groups agreed with most important treatment goals. These included improvement in survival, improvement in QoL, improvement in hematopoiesis, and transfusion independence. The L/BMT program of British Columbia added that prevention of infection, time to remission, and minimization of toxicity and adverse events associated with treatment are also important treatment goals. In addition to these objectives, the ATB-MPG noted that the reduction of burden on caregivers is also an important treatment goal. All 4 clinician groups also agreed that the currently available treatments offer short survival advantage and short transfusion independence. The short remissions often require several monthly cycles of the therapy, up to 6 cycles, to achieve maximal effect, and translate to an extended period of transfusion dependence, as noted by CLSG. CLSG also added that once a maximally achieved response is lost, disease progression is quite rapid, typically followed quickly by death due to poor salvage therapies. The ATB-MPG added that time to remission for azacitidine is around 4 months and maximal response can take more than 6 months; during this time, patients are transfusion-dependent and have significant burden of disease. The treatment under review was noted by the ATB-MPG as having a median time to response of 1.2 months (faster time to response and clinical improvement). The L/BMT program of British Columbia added that the primary unmet goals for this population are low response rates and short OS.

All 4 clinician groups also agreed that patients with AML who are not eligible for standard 7 plus 3 induction therapy (older or with comorbidities) have the greatest unmet need.

Place in Therapy

All 4 clinician groups agreed that the combination of venetoclax and azacitidine would replace current front-line therapies, including azacitidine alone in a population that has an unmet medical need. The ATB-MPG noted that some patients who received induction chemotherapy and relapsed but are no longer eligible for transplantation, or who relapsed after transplantation and had never received an HMA such as azacitidine before, would commonly use azacitidine and the combination of azacitidine and venetoclax would be expected to be more effective; however, this population was not included in the current reimbursement request.

All 4 clinician groups also agreed it would not be appropriate to recommend that patients try other treatments before initiating treatment with venetoclax and azacitidine, as this therapy is for first-line use and there is no evidence to support sequencing this combination after other treatments. The groups added that the clinical trial was also for newly diagnosed patients.

With respect to sequencing, all 4 clinician groups agreed that the combination of venetoclax and azacitidine would replace current first-line treatment. CLSG also added that after failure of the combination of venetoclax and azacitidine, possible therapeutic options would include therapy targeted to a specific molecular lesion, if present and available, an early phase clinical trial, LDAC, BSC, or palliation. L/BMT agreed with CLSG on this approach and noted that, currently, there is no standard of care practice for patients who fail first-line treatment. The ATB-MPG noted that in second-line treatment, if patients have FLT3-positive AML, they can receive gilteritinib and, if they have FLT3-negative AML, they can receive cytarabine if they are being treated with azacitidine, or with azacitidine if they are being treated with cytarabine. In third-line treatment, the ATB-MPG noted that hydroxyurea and transfusions as well as basic supportive care can be used.

Patient Population

All 4 clinician groups agreed that patients with newly diagnosed AML who are unfit for intensive chemotherapy (due to age, comorbidities, or patient decision not to undergo intensive treatment that is potentially curative) are best suited for treatment with venetoclax and azacitidine, as these patients need more effective therapy options. The L/BMT program of British Columbia noted that the age cut-off of 75 years or greater, as per the Health Canada indication, is based on the clinical trial; however, they expect that, in most centres in Canada, only a small number of patients older than 70 years of age are treated with intensive chemotherapy. The group also added there is some evidence that AML patients with mutations in genes 1DH1 or 1DH2 have a particularly good response to azacitidine and venetoclax, but patients with other genetic subsets of AML as well as de novo and secondary AML also appear to benefit.

All 4 clinician groups also agreed there is a standard diagnosis of AML (i.e., the presence of greater than 20% myeloid blasts in the bone marrow or peripheral blood). Through clinical examination and judgment, as well as bone marrow results, patients can be objectively diagnosed. The groups added that testing is widely available and that patients should be treated at the time of diagnosis, as they would be expected to decline rapidly and develop serious infections or other complications that preclude effective treatment if they are not treated at diagnosis.

It was noted that patients who are least suitable for the venetoclax and azacitidine treatment are those who are younger and fit, without significant comorbidities, or very frail older adults. The clinician groups also noted that patients who are unable to travel to outpatient clinicians to receive azacitidine would be least suitable for treatment. Also, the L/BMT program of British Columbia noted that patients who are not able to have regular blood work monitoring for tumour lysis syndrome during the initial ramp-up and regular monitoring of blood work later on are also not good candidates.

With respect to identifying patients who are most likely to exhibit a response to treatment with venetoclax and azacitidine, the clinician groups noted that some patients with specific molecular mutations (IDH1 or IDH2) on next-generation sequencing may be expected to respond better; however, this is based on subgroup analysis and there is no patient group that would not be expected to benefit within the group of patients this reimbursement request is for. The groups also noted that at this point, there is no specific test or biomarker to indicate who will or will not respond or benefit from treatment.

Assessing Response to Treatment

All 4 clinician groups agreed that important outcomes in clinical practice are remission status following treatment, tolerance of treatment, QoL, and transfusion requirements. The groups added that bone marrow biopsy to assess disease response is an important outcome; however, it was done more frequently in the clinical trial than in practice.

All 4 groups agreed that a clinically meaningful response to treatment would be remission status on bone marrow biopsy, reduced or eliminated transfusion requirements for red blood cells and platelets, and improvement in symptoms (i.e., infections, bleeding, improved functional status due to improved hemoglobin and fewer hospital admissions or outpatient visits for transfusion support). Additionally, the groups added that in patients without CR or CRi (incomplete count recovery), a partial remission or improvement in blood counts may also be a meaningful improvement for some patients.

CLSG, L/BMT program of British Columbia, and the ATB-MPG noted that response should be assessed with a bone marrow biopsy as well as evaluation of blood counts following 1 to 2 cycles (4 to 8 weeks) of azacitidine and venetoclax. The L/BMT program of British Columbia noted that in patients achieving CR or CRi, they would suggest repeating the bone marrow aspirate biopsy as clinically indicated (e.g., repeated if there is concern that a patient is losing response due to worsening blood counts or the appearance of circulating blasts). The group added that in patients with less than CR or CRi after 1 to 2 cycles who continue on treatment, they would generally repeat bone marrow biopsies every 3 to 4 months to evaluate a response. The ATB-MPG added that once remission or maximal response is obtained, repeat bone marrow biopsy would be indicated if there is clinical deterioration or significant cytopenias requiring reassessment of disease status. The OH-CCO’s Hem-DAC noted that treatment should be assessed frequently with regular CBC and bone marrow assessments, as per clinician judgment.

Discontinuing Treatment

All 4 clinician groups agreed that adverse events such as severe nausea, neutropenic infections, and severe infections may lead to a decline in patients’ ability to safely administer the treatment. The groups also agreed that failure of response or disease progression (significant increase in bone marrow blasts), treatment-related toxicities, and patient preference would be the primary reasons for treatment discontinuation.

Prescribing Conditions

All 4 clinician groups agreed that the most appropriate setting for treatment administration can be in the community setting and outpatient clinic. Inpatient hospital treatment may also be required due to tumour lysis syndrome or AML complications while continuing on the treatment. CLSG noted there should be expertise in the outpatient clinic with chemotherapy preparation and administration. The L/BMT program of British Columbia noted that treatment with venetoclax and azacitidine requires monitoring of blood counts, renal function, and electrolytes more frequently early on during the first week of administration due to the risk of tumour lysis syndrome. CLSG was of the opinion that the treatment should be given in a setting where there is blood-bank support; physician, nursing, and pharmacy expertise in chemotherapy; an ability to deliver IV fluids and antibiotics; and an ability to admit patients to hospital for complications of treatment.

Additional Considerations

OH-CCO’s Hem-DAC noted that venetoclax dose adjustment with co-administration of azole is sometimes required. In addition, the group added that in patients presenting with hyperleukocytosis, a longer ramp-up phase should be considered when initiating venetoclax.

The group added that the additional toxicities of the combination are largely related to increased myelosuppression and increased rates of febrile neutropenia early on during treatment, but this is manageable and does not offset the benefit of the treatment combination. The group also strongly supported the reimbursement of this treatment for older and unfit patients with AML due to the large, anticipated benefit for this group of patents, where there currently exists an unmet need for more effective treatment options.

CLSG noted that patients are already treated in many jurisdictions with azacitidine alone and adding an oral medication, like venetoclax, which is well tolerated with a straightforward administration schedule, does not increase the complexity of the treatment regimen. The group also added that the benefits obtained with this combination are seen much quicker than with azacitidine alone.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation.

The drug programs indicated that current treatment options for patients with newly diagnosed AML who are ineligible for intensive chemotherapy include azacitidine, LDAC, and BSC. The reimbursement of venetoclax plus azacitidine would likely replace single-agent azacitidine in this treatment setting. Azacitidine is funded in most jurisdictions for patients with AML who are ineligible for intensive chemotherapy, and some jurisdictions fund alternate dosing schedules for azacitidine (i.e., 5 to 2-2, and 6 consecutive days) in addition to the schedule of 7 consecutive days. However, it was noted that some patients 75 years of age and older may be fit to tolerate intensive chemotherapy. The ramp-up dosing schedule for venetoclax with azacitidine differs significantly from the ramp-up dosing schedule already in use for CLL indications, and the current packaging for venetoclax is designed for the CLL ramp-up dosing schedule. Venetoclax plus azacitidine includes an oral and an IV and subcutaneous drug and, therefore, would be reimbursed through different programs in some jurisdictions. The drug programs identified the potential for indication creep for patients with a high risk of MDS, those who have progressed or have had an inadequate response on low-dose chemotherapy for AML, and patients who have relapsed after induction chemotherapy and are not eligible for stem cell transplant and who are then treated with azacitidine. It was noted that treatment combination may require the need for increased health care resources (i.e., hospital admission and additional pharmacy and nursing resources for the potential management of tumour lysis syndrome and monitoring for drug interactions). Affordability was also identified as an issue since the combination is expected to replace azacitidine monotherapy.

The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Clinical Evidence

The clinical evidence included in the review of venetoclax (Venclexta) in combination with azacitidine is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. The second section includes indirect evidence from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of venetoclax in combination with azacitidine for the treatment of patients with newly diagnosed AML who are 75 years or older, or who have comorbidities that preclude use of intensive induction therapy.

Methods

Studies selected for inclusion in the systematic review will include pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist (https://www.cadth.ca/resources/finding-evidence/press).²¹

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) through Ovid and Embase (1974–) through Ovid. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were Venclexta (venetoclax) and AML. Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. See Appendix 1 for the detailed search strategies.

The initial search was completed on February 11, 2021. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC) on June 10, 2021.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist (https://www.cadth.ca/grey-matters).²² Included in this search were the websites of regulatory agencies (US FDA and European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented through contacts with appropriate experts. In addition, the sponsor of the drug was contacted for information regarding unpublished studies.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

Findings From the Literature

One study²³ was identified from the literature for inclusion in the systematic review (Figure 1). The included study is summarized in Table 6. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

VIALE-A

VIALE-A is a phase III, randomized, double-blind, placebo-controlled, multi-centre study comparing venetoclax plus azacitidine with placebo plus azacitidine in adults with newly diagnosed AML aged 18 years or older and ineligible for standard induction therapy due to age or comorbidities. The trial was conducted at 134 sites in Europe, Asia, South America, Canada, and the US. Trial characteristics are summarized in Table 6.

The primary objective was to evaluate whether venetoclax plus azacitidine would improve OS and composite complete remission rate (CR + CRi) compared with placebo plus azacitidine.

The secondary objectives were to evaluate whether, compared with placebo plus azacitidine, venetoclax plus azacitidine would improve the CR rate, CR + CRh rate, CR + CRi rate by the initiation of cycle 2, transfusion-independence rate, MRD response rate, response rates and OS in molecular subgroups, fatigue, GHS/QoL, and EFS. Improvement of fatigue and GHS/QoL were measured using patient-reported outcome assessments (Patient-Reported Outcomes Measurement Information System Short Form v1.0–Fatigue 7a [PROMIS 7a] and European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 [EORTC QLQ-C30]). Definitions and further details of outcomes are given in Table 8.

Exploratory objectives included the study of biomarkers predictive of venetoclax plus azacitidine activity and the evaluation of additional subscales and items from the EORTC QLQ-C30 and EuroQol 5-Dimensions 5-Levels questionnaire (EQ-5D-5L) scales.

A total of 579 patients were screened and 433 patients were randomized in a 2:1 ratio to venetoclax plus azacitidine (287 patients), or placebo plus azacitidine (146 patients). Randomization was stratified for age (18 years to < 75 years, ≥ 75 years) and cytogenetic risk (intermediate, poor) for protocol amendment 1 and subsequent amendments. Prior to amendment 1, stratification was by age and region; the 2 patients (1 in each group) randomized under the original protocol were not included in the efficacy analysis. An additional open-label cohort of up to 12 patients was recruited to receive venetoclax plus azacitidine so as to provide pharmacokinetic and safety data required by the Chinese regulatory authorities before the recruitment of Chinese patients into the double-blind randomized portion of the study. These patients were not included in the efficacy or safety analyses for the main study and will not be described further in this report.

The first patient was randomized on February 6, 2017 and the last on May 31, 2019. Analysis of the co-primary efficacy end point of CR + CRi was planned for 6 months after the first 225 patients were randomized (first interim analysis [IA1]), at which time an interim analysis of survival was conducted. Interim data were reviewed by an independent data monitoring committee (IDMC) and the trial proceeded blinded. A second interim analysis (IA2) was planned when approximately 270 OS events accrued, before the final analysis at approximately 360 OS events. Review of the results against preplanned stopping criteria led to a recommendation by the IDMC to stop the trial on March 16, 2020. Data cut-off for IA1 was October 1, 2018 and January 4, 2020 for IA2.

Populations

Inclusion and Exclusion Criteria

Eligible patients had to be 18 years or older with confirmed AML according to WHO criteria and be considered ineligible for induction therapy on account of age (≥ 75 years) or significant cardiac, pulmonary, renal, hepatic, or other comorbidity. Patients aged 75 years and older had to have an ECOG PS of 0 to 2, and those aged 18 to 74 years could have an ECOG 0 to 3.

Patients were ineligible if they had acute promyelocytic leukemia, AML with favourable risk cytogenetics such as t(8;21), inv(16), t(16;16) or t(15;17) (as per NCCN Guidelines Version 2, 2016 for AML), AML with known active CNS involvement, a history of myeloproliferative neoplasm, or if they had previously received an HMA and/or any chemotherapeutic agent for MDS, or CAR T-cell therapy or any other experimental therapy.

Baseline Characteristics

Table 7 shows the baseline characteristics for the efficacy population. The mean age overall was 75.4 (SD = 5.95) years, and 60.6% of patients were aged 75 years or older. The majority were male (60.1%) and White (75.6%), followed by Asian (23.0%). Most had an ECOG PS of 1 (41.5%) or 2 (29.9%). The most common reasons patients were ineligible for standard induction therapy were because they were aged 75 years or older (56.8%) and had an ECOG PS of 2 or 3 (43.2%). Most patients (75.2%) had de novo AML; of those with secondary AML, 67.3% had prior MDS or chronic myelomonocytic leukemia. Cytogenetics were intermediate risk in 62.9% and poor in 37.1%. Of individual mutations, 23.9% of patients with known mutation status had mutations detected in IDH1 and/or IDH2, 16.2% had FLT3-ITD mutations and/or tyrosine kinase domain (TKD) mutations, 17.7% had nucleophosmin 1 (NPM1) mutations, and 20.9% had tumour protein p53 (TP53) mutations. Half (49.9%) had a bone marrow blast count of 50% or greater, 21.8% had a blast count between 30% and less than 50%, and 29.2% had a blast count of less than 30%. More than half (54.8%) had required a red blood cell or platelet transfusion within 8 weeks of first study treatment.

The demographic and baseline characteristics were well balanced. A greater proportion of patients receiving placebo plus azacitidine had mutations in FLT3 (20.4%) compared with those receiving venetoclax plus azacitidine (14.1%); conversely, a greater proportion of patients receiving venetoclax plus azacitidine had mutations in TP53 (23.3%) compared with placebo plus azacitidine (16.3%). Overall cytogenetic risk was a stratification variable and was well balanced.

Interventions

Venetoclax in the venetoclax plus azacitidine group was dosed at 400 mg oral per day after the initial dose titration of 100 mg at cycle 1 day 1, 200 mg at day 2, and 400 mg at day 3 and thereafter. Dosing was based on a phase Ib dose-escalation study of venetoclax in combination with HMAs (azacitidine and decitabine) in treatment-naive patients with AML who were aged 65 years or older and ineligible for the standard induction regime (Study M14 to 358).

Azacitidine in the venetoclax plus azacitidine and azacitidine groups was dosed at 75 mg/m² subcutaneous or IV (depending on local practice) on days 1 to 7 of a 28-day cycle. The dose is as specified for the treatment of adult patients with AML in the EU Summary of Product Characteristics and for MDS in the US dose prescribing information.

Treatment was planned for a minimum of 6 cycles. Treatment could continue as long as the patient derived clinical benefit and did not have documented disease progression or develop unacceptable toxicity.

To allow for hematologic recovery from cytopenias, the dosing of venetoclax or placebo could be interrupted, the start of the next cycle delayed, or the duration reduced according to preplanned criteria. If the recovery did not reach a certain threshold, then the administered dose of azacitidine could be reduced by preplanned steps.

• If a patient achieved CRi or a morphologic leukemia-free state (MLFS) after cycle 1, they could interrupt venetoclax or placebo dosing from day 29 until an absolute neutrophil count (ANC) of 500/μL or greater, or for up to 14 days. Cycle 2 administration would be delayed until ANC was 500/μL or greater.

• If a patient experienced a new onset grade 4 neutropenia for more than 1 week during subsequent cycles that was not thought to be due to the underlying disease, venetoclax or placebo dosing could be interrupted until ANC was 500/μL or greater.

• After cycle 3, a patient in CR/CRi who needed interruption or delay of study drug for cytopenia could receive venetoclax plus azacitidine for 21 out of 28 days of each cycle.

• If a patient showed hematological recovery (ANC or platelets) within 14 days after completion of a cycle, the duration of venetoclax was reduced to 21 days of the cycle. During subsequent cycles, if hematologic recovery with more than a 25% increase above the nadir was not seen within 21 days after cycle completion, the azacitidine dose was reduced to 50% in patients with bone marrow cellularity of 15% to 50%, and to 33% in patients with bone marrow cellularity of less than 15%.

Protocol-specified dose adjustments were also made to adjust for drug interactions with systemic anti-fungal agents and other drugs that produced moderate or strong inhibition of CYP3A. Venetoclax or placebo was to be reduced at least twofold in patients receiving moderate CYP3A inhibitors and at least eightfold in those receiving strong CYP3A inhibitors. Moderate CYP3A inducers were excluded during the ramp-up phase and used with caution and after discussion with the sponsor.

All patients received prophylaxis against tumour lysis syndrome with oral and/or IV hydration and uric acid reducer and were admitted for monitoring during ramp-up of venetoclax plus placebo dosing. Patients with a white blood cell count greater than 25 × 10⁹/L received cytoreduction before dosing.

Outcomes

Table 8 provides definitions of the efficacy end points that were assessed in the clinical trials included in this review and identified in the CADTH review protocol. Response measures (CR, CRi), partial remission, MLFS, resistant disease, and morphologic relapse were based on the revised guidelines for the International Working Group (IWG) for AML. Progressive disease was defined according to European LeukemiaNet recommendations. CRh is a derived end point based on bone marrow and hematological measurements. These end points are further summarized in Table 8. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

EORTC QLQ-C30 (Version 3.0) is a cancer-specific measure of health-related quality of life (HRQoL).²⁴ It comprises 30 individual questions organized into 5 functional scales (physical, role, cognitive, emotional, and social function), 3 symptom scales (fatigue, pain, and nausea and vomiting), 6 single-item symptom scales (dyspnea, insomnia, appetite loss, constipation, diarrhea, financial impact), and a global QoL scale (GHS/QoL). Function and symptoms are assessed over a 1-week recall period. Most questions have 4 response options (“not at all,” “a little,” “quite a bit,” “very much”), with scores on these items ranging from 1 to 4.²⁵ For the 2 items that form the GHS/QoL scale, the response format is a 7-point Likert-type scale, with anchors between 1 (very poor) and 7 (excellent).²⁵ Each raw scale score is converted to a standardized score ranging from 0 to 100, with a higher score reflecting better function on the function scales, higher symptoms on the symptom scales, and better QoL on the QoL scales.

Construct validity and internal consistency reliability were assessed by 2 separate studies using convenience samples of cancer patients in Singapore²⁶ and Kenya,²⁷ most of whom had breast or colon cancer. No relevant studies of responsiveness were found. Two studies estimated the minimal important difference (MID), 1 in patients with breast and small-cell lung cancer, using an anchor-based approach using change as measured by the subjective significance questionnaire,²⁸ and the other in Canadian patients newly diagnosed with breast and colorectal cancer.²⁹ Both estimated an MID of 10 points. A third study in Canadian patients³⁰ used the EORTC (the GHS/QoL questions) to estimate MIDs for the overall EORTC QLQ C30 scales, but the method did not allow estimation of the MID for the GHS/QoL subscale itself.

The PROMIS 7a is a 7-item, patient-reported, tool that measures both the experience of fatigue and the interference of fatigue on daily activities over the past week. Responses are measured on a 5-point Likert scales from 1 = never to 5 = always, with total scores ranging from 7 to 35, with higher scores indicating greater fatigue.

There were no published validation studies of PROMIS 7a in cancer. Concurrent and discriminant validity were assessed in a mixed group of non-cancer patients and healthy controls. Known-groups validity was assessed by comparing PROMIS 7a scores in the clinical samples with healthy controls.³¹

The researchers who conducted the VIALE-A trial assessed the MID using anchor- and distribution-based approaches in a group of AML patients from the VIALE-C study.¹ A 3-point difference that fell within the range of 3 to 5 proposed in the literature was considered an appropriate MID for patients with AML.

The EQ-5D-5L is a generic HRQoL instrument applicable to a wide range of health conditions.^32,33 The first of 2 parts of the EuroQol 5-Dimensions questionnaire (EQ-5D) is a descriptive system that classifies respondents (aged ≥ 12 years) in 5 dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. The EQ-5D-5L has 5 possible domains, representing “no problems,” “slight problems,” “moderate problems,” “severe problems” and “extreme problems.” Respondents are asked to choose the level that reflects their health state for each of the 5 dimensions, corresponding with 3,125 different health states. The second part is a visual analogue scale in which patients are asked to appraise their overall QoL on a scale of 0 to 100. There are 3 outputs: a 5-digit profile indicating the health states on the 5 dimensions, a population preference-weighted health index score based on the descriptive system, and a self-reported assessment of health status based on the EuroQol Visual Analogue Scale (EQ VAS).

Validity was assessed in 184 Canadian cancer patients with breast, colorectal, or lung cancer.³⁴ The EQ-5D was able to discriminate between groups based on self-reported health status (excellent/good versus fair/very poor), somewhat based on ECOG PS (0 versus 1 to 3), but not for stage of cancer.³⁴ Internal consistency reliability was calculated for the same study, and all 5 functioning scales along with GHS showed acceptable consistency (alpha > 0.70). Responsiveness was not reported. No relevant studies reported the MID.

Statistical Analysis

The analysis of the co-primary efficacy end point of CR + CRi was planned for 6 months after the first 225 patients were randomized. Three analyses were planned for the co-primary efficacy end point of OS:

• at the same time as the CR + CRi analysis

• when approximately 270 OS events (75%) had occurred of a planned total 360 events (IA2)

• when approximately 360 events had occurred (final analysis)

There was a formal interim analysis for safety after approximately 20 patients received study medication and had been followed for at least 3 months. Subsequently, safety reviews were conducted every 3 months.

Interim data were reviewed by an IDMC that could make recommendations for the ongoing conduct of the trial. The trial design allowed for early regulatory submission in the EU following IA1, if desired. Following review of the results of IA, the IDMC recommended no modification and the trial continued without unblinding. Review of the results of IA2 led to an IDMC recommendation to stop the trial on March 16, 2020.

Table 9 summarizes the methods used for statistical analysis of the efficacy end points. Time-to-event end points were analyzed with stratified Kaplan–Meier estimates and stratified log-rank comparisons. Proportions were compared by calculation of stratified ORs with statistical comparison by stratified Cochrane-Mantel-Haenszel tests. Stratification was by age and cytogenetic risk categories. Patient-reported outcomes were compared by linear mixed-effects regression models fitted to longitudinal data with covariance structure.

The sample size calculation assumed that patients would be randomized 2:1 to venetoclax plus azacitidine to placebo plus azacitidine. Other assumptions were:

• A 2-sided alpha of 0.05 would be divided between co-primary end points to allocate 0.01 alpha to CR + CRi and 0.04 alpha to OS.

• The rate of CR + CRi was projected to be 28% for placebo plus azacitidine and 55% for venetoclax plus azacitidine.

• A median OS of 10.4 months was projected for placebo plus azacitidine and 14.9 months (HR = 0.7) for venetoclax plus azacitidine.

• An interim OS 75% calculation with O’Brien-Fleming boundary, with an interim data cut-off at 270 deaths.

Based on these assumptions, 225 patients (150 venetoclax plus azacitidine, 75 placebo plus azacitidine) would give 88% power to detect difference at a 2-sided alpha of 0.01, and 360 death events would have 86.7% power to detect a difference in OS with a 2-sided alpha of 0.04. Approximately 400 patients would be randomized: 267 to venetoclax plus azacitidine, and 133 to placebo plus azacitidine.

A hierarchical testing strategy was used for the control of multiplicity. The co-primary analysis of CR + CRi was performed on the first 225 patients (IA1). Two interim analyses of OS were planned, the first at IA1, with an administrative significance level of 0.0001. IA2 was performed after approximately 270 deaths (75% of predetermined deaths), using a Lan-DeMets alpha-spending function with O’Brien-Fleming boundary to control the 1-sided false-positive rate. Based on this analysis and pre-specified stopping rules, the IDMC made a recommendation to stop for success at IA2 rather than proceed to the final analysis. For the CR + CRi rate, CR + CRh rate, CR + CRi rate by the initiation of cycle 2, CR + CRh rate by the initiation of cycle 2, CR rate, and transfusion independence rate outcomes, the information fraction to be used at IA2 was calculated based on the proportion of the planned number of patients who had reached the desired amount of follow-up (e.g., for CR + CRi, at least 6 months of follow-up).

An overall 2-sided significance level of 0.05 was initially allocated between the co-primary end points: 0.01 to the analysis of CR + CRi and 0.04 to the analysis of OS. If the results of statistical testing for CR + CRi at IA1 were significant, then the 0.01 allocated to CR + CRi would be recycled to the OS analysis. If the results of statistical testing for OS were significant, then the fixed testing procedure would be performed at a 2-sided significance level of 0.05 for each of the selected secondary efficacy end points in sequence. If the results of statistical testing for OS were not significant, then statistical significance would not be declared for any of the end points.

Table 10 shows the actual alpha-spending boundary and information fraction for the end points tested under the hierarchical testing strategy, in order of testing.

Drop-outs or missing data were handled using the following rules:

• Patients who had not died were censored at the last known dates they were known to be alive on or before the cut-off date for the analysis of OS.

• Patients who were randomized but did not have an IWG disease assessment were considered nonresponders in the calculation of CR, CR + CRi.

• Patients who did not receive the study drug were considered post-baseline transfusion-dependent in the analysis of post-baseline transfusion-independent rates.

• Patients who did not experience relapse or death after response were censored at the date of last disease assessment (bone marrow or hematology laboratory measurement).

• Patients who did not experience an EFS event and did not start on post-treatment therapy were censored at the time of the last disease assessment date on or before the data cut-off date. Patients who did not experience an EFS event and started on post-treatment therapy were censored at the time of initiation of post-treatment therapy.

• There was no imputation of PROMIS 7a scores at a time point if scores were missing entirely; the summary was of available data. The scoring of missing individual items was according to the manual.

• There was no imputation of EORTC QLQ-C30 and subscales data at a time point if scores were missing entirely; the summary was of available data. If there were missing items for a scale (i.e., the participant did not provide a response), the score for the scale could still be computed if there were responses for at least 1-half of the items. In calculating the scale score, the missing items were simply ignored — an approach that assumed the missing items had values equal to the average of those items that the respondent completed.²⁵

Analysis Populations

Two analysis populations were identified:

• The efficacy population included all patients randomized under protocol amendment 1 and subsequent amendments. It excluded the 2 patients randomized under the original protocol and the open-label China cohort of 10 patients (n = 431).

• The safety population included all patients under the study protocol who received at least 1 dose of the study drug, but not the open-label China cohort (n = 427).

The longitudinal analysis population in the supplementary analysis of patient-reported outcomes included all patients in the efficacy population who survived up to a given time point and had available data for at least 1 patient-reported outcome measurement at baseline and at that time point.

Results

Patient Disposition

Table 11 shows the patient disposition. Of the 579 patients screened, 2 patients were randomized under the original protocol (group 1) and 431 patients were randomized under protocol amendment 1 and subsequent amendments (group 2). As the stratification variables changed under amendment 1, only patients in group 2 were included in the efficacy population. All eligible patients were included in the safety population.

A total of 579 patients was screened, with 146 excluded on screening, primarily because they did not meet the inclusion or exclusion criteria. Of the 431 patients randomized under amendment 1 and subsequent amendments, 286 were randomized to venetoclax plus azacitidine and 144 to placebo plus azacitidine. The primary reason for study discontinuation was death, in 56.7% of patients randomized to venetoclax plus azacitidine, and 75.2% of patients randomized to placebo plus azacitidine. Both loss to follow-up and withdrawal of consent were very low: less than 2% were lost to follow-up in both arms, and less than 2.5% withdrew consent. Three patients randomized to venetoclax plus azacitidine and 1 randomized to placebo plus azacitidine did not receive their assigned treatment and were excluded from the safety population.

Exposure to Study Treatments

Table 12 shows a summary of exposure for the safety population. Patients randomized to venetoclax plus azacitidine had longer exposure on average than those randomized to placebo plus azacitidine. The median duration of exposure for patients randomized to venetoclax plus azacitidine was 7.6 months (range, 0.0 to 30.7) and the median number of cycles was 7 (range, 1 to 30), compared with 4.3 months (range, 0.1 to 24.0) and 4.5 cycles (range, 1 to 26), respectively, for placebo plus azacitidine. A total of 106 patients (37.4%) who received venetoclax plus azacitidine and 42 (29.1%) patients who received placebo plus azacitidine received 10 or more cycles.

More patients who received venetoclax plus azacitidine had a dose reduction and/or dose interruption than those who received placebo plus azacitidine, either drug in each combination, venetoclax or placebo alone, and azacitidine alone. Of the patients who received venetoclax plus azacitidine, 21.6% had 1 dose reduction, 4.6% had 2 dose reductions, and 2.5% had more than 2 dose reductions, compared with 18.1%, 2.1%, and 2.1%, respectively, for patients who received placebo plus azacitidine. Of the patients who received venetoclax plus azacitidine, 94.3% had at least 1 dose interruption for any reason and 66.1% had at least 1 dose interruption for count recovery, compared with 77.8% and 20.1%, respectively, for patients who received placebo plus azacitidine. Of the patients who received venetoclax plus azacitidine, 52.8% had a dose interruption of venetoclax or placebo and 65.3% had a dose interruption of azacitidine, compared with 27.7% and 25.5%, respectively, for patients who received placebo plus azacitidine.

Efficacy

Table 13 shows an overall summary of the efficacy outcomes and the subgroup analyses for the efficacy population at the time of the second interim analysis. Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported subsequently. Between-treatment differences that were statistically significant with controlled multiplicity (co-primary end points and sequential testing protocol) are indicated with a footnote.

At the time of the second interim analysis, the median duration of follow-up for patients randomized to venetoclax plus azacitidine was 20.7 months (95% CI, 20.1 to 22.0), and for those randomized to placebo plus azacitidine, it was 20.2 months (95% CI, 19.6 to 22.4).

Overall Survival

Table 13 summarizes the OS results for the efficacy population at IA2 (data cut-off: January 4, 2020) and Figure 2 shows the Kaplan–Meier survival curves.

Venetoclax plus azacitidine improved survival over placebo plus azacitidine (P value [stratified log-rank test] < 0.001; P value boundary = 0.02), with median survival in patients randomized to venetoclax plus azacitidine of 14.7 months (95% CI, 11.9 to 18.7) compared with 9.6 months (59% CI, 7.4 to 12.7) in patients randomized to placebo plus azacitidine (HR for mortality = 0.662; 95% CI, 0.518 to 0.845). The estimated survival at 12 months for venetoclax plus azacitidine was 55.8% (95% CI, 49.7% to 61.5%); for placebo plus azacitidine it was 43.8% (95% CI, 35.5% to 51.8%).

A sensitivity analysis that included all of the data in the extracted database showed similar results, with an HR for mortality of 0.653 (95% CI, 0.513 to 0.832), and an estimated survival rate at 12 months of 56% for venetoclax plus azacitidine (95% CI, 49.9% to 61.6%) and 43.9% for placebo plus azacitidine (95% CI, 35.6% to 51.9%).

Figure 2: Overall Survival, Efficacy Population, IA2

CI = confidence interval; COX PH = Cox proportional hazards model; IA2 = second interim analysis; PBO + AZA = placebo plus azacitidine; VEN + AZA = venetoclax plus azacitidine.

Note: Data cut-off was January 4, 2020.

Source: Clinical Study Report.¹

Event-Free Survival and Duration of Event-Free Survival

Table 13 summarizes the results for EFS for the efficacy population at IA2 (data cut-off: January 4, 2020) and Figure 3 shows the Kaplan–Meier survival curves.

Venetoclax plus azacitidine improved EFS over placebo plus azacitidine (P value [stratified log-rank test] < 0.001; P value boundary of 0.032), with a median EFS duration of 9.8 months (95% CI, 8.4 to 11.8) for patients randomized to venetoclax plus azacitidine compared with 7.0 months (95% CI, 5.6 to 9.5) for patients randomized to placebo plus azacitidine (HR for EFS = 0.632; 95% CI, 0.502 to 0.796). At 12 months, 43.5% of the patients randomized to venetoclax plus azacitidine were free of confirmed morphologic relapse or disease progression, treatment failure, or death, compared with 31.3% of patients randomized to placebo plus azacitidine. A higher proportion of patients in the venetoclax plus azacitidine group (43.5%) had a confirmed morphologic relapse or disease progression event than in the placebo plus azacitidine group (28.7%); conversely, a higher proportion of patients had death as an event in the placebo plus azacitidine group (61.5%) than in the venetoclax plus azacitidine group (54.5%).

Figure 3: Event-Free Survival, Efficacy Population, IA2

CI = confidence interval, COX PH = Cox proportional hazards model, IA2 = second interim analysis; PBO + AZA = placebo plus azacitidine, VEN + AZA = venetoclax plus azacitidine.

Note: Data cut-off was January 4, 2020.

Source: Clinical Study Report.¹

Response: CR + CRi, CR, CR + CRh

Table 13 summarizes the results for composite complete remission (CR + CRi) for the efficacy population at the second interim assessment. The table also summarizes the results for CR + CRh, which has required minimum values for neutrophils and platelet counts.

CR + CRi at IA1 was a co-primary efficacy end point, as assessed for the first 226 patients (data cut-off: October 1, 2018). Venetoclax plus azacitidine improved CR + CRi (P value for stratified Cochran-Mantel-Haenszel [CMH] test < 0.001; P value boundary = 0.01) with 65.3% (95% CI, 57.0 to 73.0%) of patients meeting the end point compared with 25.3% (95% CI, 16.2% to 36.4%) of those who received placebo plus azacitidine.

Similar results were seen for conventional care regimen at IA2, with 66.4% (95% CI, 60.6 to 71.9%) of patients randomized to venetoclax plus azacitidine reaching the end point of CR + CRi compared with 28.3% (95% CI, 21.1% to 36.3%) of patients randomized to placebo plus azacitidine. This end point was not tested. Venetoclax plus azacitidine improved CR (P value [stratified CMH test] < 0.001; P value boundary = 0.047), with 36.7% (95% CI, 31.1% to 42.6%) of patients meeting the criteria for CR compared with 17.9% (95% CI, 12.1% to 25.2) of patients randomized to placebo plus azacitidine.

Among patients who had a partial response, 1.0% of those randomized to venetoclax plus azacitidine and 2.1% of those randomized to placebo plus azacitidine met the criteria for partial remission, while 8.4% of those randomized to venetoclax plus azacitidine and 4.1% of those randomized to placebo plus azacitidine met the criteria for MLFS.

Venetoclax plus azacitidine also improved early composite complete remission (P value [stratified CMH test] < 0.001; P value boundary = 0.05), with 43.4% (95% CI, 37.5% to 49.3%) of patients reaching the end point of CR + CRi by the beginning of cycle 2, compared with 7.6% (95% CI, 3.8 to 13.2) of patients randomized to placebo plus azacitidine. Results for time to first response (next section) were consistent.

Results for CR + CRh were consistent with those for composite complete remission, with 64.7% (95% CI, 58.8% to 70.2%) of patients randomized to venetoclax plus azacitidine reaching the end point of CR + CRh compared with 22.8% (95% CI, 16.2% to 30.5%) of patients randomized to placebo plus azacitidine, and similar results for time to first and time to best response.

Time to Response

Table 13 shows time to first and time to best response for CR + CRi and CR + CRh. Time to first response was 1.3 months (95% CI, 0.6 to 9.9) for patients randomized to venetoclax plus azacitidine and 2.8 months (95% CI, 0.8 to 13.2) for those randomized to placebo plus azacitidine. Median times to best CR or CRi were 2.3 (95% CI, 0.6 to 24.5) months and 3.7 (0.8 to 13.2) months for patients randomized to venetoclax plus azacitidine and placebo plus azacitidine, respectively. Very similar results were seen for CR + CRh.

Duration of Response

Table 13 shows the duration of response for CR + CRi and CR for the efficacy population at IA2 (data cut-off: January 4, 2020). The duration of response for patients who reached CR or CRi was 17.5 months (95% CI, 13.6 to not estimable) for patients randomized to venetoclax plus azacitidine and 13.4 (95% CI, 8.5 to 17.6) months for patients randomized to placebo plus azacitidine. At 12 months, 60.6% of responding patients randomized to venetoclax plus azacitidine still met the response criteria compared with 51.0% of responding patients randomized to placebo plus azacitidine.

Post-Baseline Transfusion Independence

Table 13 summarizes independence and duration of independence from red blood cell and platelet transfusions and from both combined for the efficacy population at IA2 (January 4, 2020). The table also shows results for patients who were and were not transfusion-independent at baseline.

Venetoclax plus azacitidine improved post-baseline transfusion independence for red blood cells (P value [stratified CMH test] < 0.001; P value boundary = 0.047). For red blood cells, 59.8% (95% CI, 53.9% to 65.5%) of patients randomized to venetoclax plus azacitidine were transfusion-independent compared with 35.2% (95% CI, 27.4% to 43.5%) of patients randomized to placebo plus azacitidine, a treatment difference of 24.6% (95% CI, 15.0 to 34.2). The median duration of red blood cells transfusion independence was 199 days (range of 57 to 933 days) for patients randomized to venetoclax plus azacitidine and 193 days (range of 56 to 727 days) for patients randomized to placebo plus azacitidine. For both groups, a greater proportion of patients who were transfusion-independent at baseline were transfusion-independent post baseline compared with patients who were not transfusion-dependent at baseline.

Venetoclax plus azacitidine improved post-baseline transfusion independence for platelets (P value [stratified CMH test] < 0.001; P value boundary = 0.047). For platelets, 68.5% (95% CI, 62.8% to 73.9%) of patients randomized to venetoclax plus azacitidine were transfusion-independent compared with 49.7% (95% CI, 41.3% to 58.1%), a treatment difference of 18.9% (95% CI, 9.1% to 28.6%). The median duration of platelet transfusion independence was 210 days (range of 56 to 933 days) for patients randomized to venetoclax plus azacitidine, and 227.5 days (range of 58 to 730 days) for patients randomized to placebo plus azacitidine. For both groups, a greater proportion of patients who were transfusion-independent at baseline were still transfusion-independent post baseline compared with patients who were not transfusion-independent at baseline.

Hospitalization

Overall rates of hospitalization were not reported. A higher percentage of patients who received venetoclax plus azacitidine experienced adverse events leading to hospitalization than those who received placebo plus azacitidine (80.6% versus 66.7%). The most common adverse events leading to hospitalization involved cytopenias or were infectious: febrile neutropenia (29.3% for venetoclax plus azacitidine versus 10.4% for placebo plus azacitidine), pneumonia (14.8% versus 20.1%), anemia (4.9% versus 4.2%), neutropenia (4.6% versus 2.1%), sepsis (4.6% versus 6.9%), and thrombocytopenia (3.9% versus 0.7%).

Overall Survival and CR + CRi in Subgroups

Table 38 (Appendix 3) shows the results for the comparison of the molecular subgroups in the efficacy population, as of IA2, for OS and CR + CRi. Data were available for the subgroups of age (age < 75 years, age ≥ 75 years), ECOG PS (ECOG < 2, ECOG ≥ 2), cytogenetic risk (intermediate versus poor), de novo versus secondary AML, AML with and without myelodysplastic syndrome, blast count at baseline (< 30%, 30% to < 50%, and 50%), and mutations (IDH1 and/or IDH2, FLT3, NPM1, and TP53).

OS in patients with IDH1 and/or IDH2 and in patients with FLT3, and CR + CRi with IDH1 and/or IDH2 and in patients with FLT3 were included as end points in the sequential testing strategy.

Venetoclax plus azacitidine improved OS in patients with an IDH1 and/or IDH2 mutation compared with placebo plus azacitidine (P value unstratified log-rank test < 0.0001; P value boundary = 0.0002). Median survival for patients randomized to venetoclax plus azacitidine was not estimable, compared with 6.2 months (95% CI, 2.3 to 12.7) for patients randomized to placebo plus azacitidine (HR = 0.345; 95% CI, 0.199 to 0.598).

No statistically significant difference was identified for OS in patients with FLT3 mutation between patients randomized to venetoclax plus azacitidine and those randomized to placebo plus azacitidine (P value [unstratified log-rank test] < 0.2054; P value boundary = 0.0002). Median survival for patients randomized to venetoclax plus azacitidine was 12.7 months (95% CI, 7.3 to 23.5) compared with 8.9 months (95% CI, 5.9 to 14.7) for patients randomized to placebo plus azacitidine (HR = 0.664; 95% CI, 0.351 to 1.257).

Venetoclax plus azacitidine improved CR + CRi in patients with IDH1 and/or IDH2 mutation compared with placebo plus azacitidine (P value [Fisher’s exact test] < 0.001; P value boundary = 0.05). The composite remission rate for patients randomized to venetoclax plus azacitidine was 75.4% (95% CI, 52.9% to 87.3%), compared with 10.7% (95% CI, 2.3% to 28.2%) for patients randomized to placebo plus azacitidine.

Venetoclax plus azacitidine improved CR + CRi in patients with FLT3 mutation, compared with placebo plus azacitidine (P value [Fisher’s exact test] < 0.021; P value = 0.05). Median survival for patients randomized to venetoclax plus azacitidine was 72.4% (95% CI, 52.8% and 87.3%) compared with 36.4% (95% CI, 17.2% to 59.3%) for patients randomized to placebo plus azacitidine.

For the other subgroups, the greatest difference in point estimates for OS and CR + CRi was observed for age and cytogenetic risk. In patients aged less than 75 years, the HR for OS for the comparison of venetoclax plus azacitidine to placebo plus azacitidine was 0.888 (95% CI, 0.591 to 1.33), compared with an HR of 0.535 (95% CI, 0.394 to 0.727) in patients aged 75 years and older. In patients aged less than 75 years, the risk difference in CR + CRi for the comparison of venetoclax plus azacitidine to placebo plus azacitidine was 21.12% (95% CI, 5.6% to 36.6%), compared with 49.43% (95% CI, 38.6% to 60.2%) in patients aged 75 years and older.

In patients with intermediate cytogenetic risk, HR for OS for the comparison of venetoclax plus azacitidine versus placebo plus azacitidine was 0.566 (95% CI, 0.407 to 0.786), compared with an HR of 0.775 (95% CI, 0.538 to 1.117) in patients with poor cytogenetic risk. In patients with intermediate risk, the risk difference in CR + CRi for the comparison of venetoclax plus azacitidine versus placebo plus azacitidine was 42.72% (95% CI, 31.2% to 54.3%), compared with 29.67% (95% CI, 15.0% to 44.3%) in patients with poor cytogenetic risk.

Differences in point estimates were minimal for the other identified subgroups.

Patient-Reported Outcomes

Overall QoL was captured by the EORTC QLQ-C30 GHS/QoL and health utility was captured by the EQ-5D. Fatigue was reported in the PROMIS 7a scale and the fatigue subscale in the EORTC QLQ-C30. Nausea and vomiting, pain, dyspnea, insomnia, appetite loss, constipation, and diarrhea were all reported as part of the EORTC.

The EORTC QLQ-30 GHS/QoL and PROMIS 7a were secondary end points and included in the statistical testing strategy. EQ-5D was reported as an exploratory end point. Exploratory analyses were conducted of the time to deterioration of individual scores and subgroups.

EORTC QLQ-C30 GHS/QoL

Table 38 shows the mean scores at baseline and day 1 of subsequent treatment cycles, the least squares (LS) mean change from baseline, and the LS mean difference between venetoclax plus azacitidine and placebo plus azacitidine, for EORTC QLQ-C30 GHS/QoL. Data are shown up to cycle 13, after which the attrition of patients due to death and discontinuation due to disease progression had reduced the study cohorts to a small number.

Figure 4 shows the time to deterioration on the EORTC QLQ-C30 GHS/QoL. At baseline, compliance in responding to the questionnaire was 92.9% and 90.9% in the venetoclax plus azacitidine and placebo plus azacitidine groups, respectively. In subsequent cycles (up to cycle 13), compliance in individual treatment groups ranged from 72.4% to 83.9%.

Figure 4: Time to Deterioration on the EORTC QLQ-C30 GHS/QoL, Longitudinal Analysis Population

AZA = azacitidine; CI = confidence interval; GHS/QoL = global health status quality of life scale; HRQoL = health-related quality of life; LAP = longitudinal analysis population; EORTC QLQ-30C = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; PBO = placebo; PRO = patient-reported outcome; VEN = venetoclax.

Source: Patient-reported outcome report.¹

The mean EORTC QLQ-C30 GHS/QoL at baseline was similar in patients in the venetoclax plus azacitidine group and placebo plus azacitidine group: 52.61 (n = 262) and 55.96 (n = 130), respectively. There was a greater change from baseline in the venetoclax plus azacitidine group than in the placebo plus azacitidine group at all points except cycle 19. The difference met or exceeded the MID of 5 points at cycles 5 and 21. There were no clinically meaningful differences in mean change from baseline between treatment groups. This end point was part of the statistical testing hierarchy, but testing failed for an end point before this. Thus, conclusions cannot be drawn based on the results of this outcome without risk of increased type I error.

The median time to deterioration for patients randomized to venetoclax plus azacitidine was 16.5 months, compared with 9.3 months for patients randomized to placebo plus azacitidine. The adjusted HR was 0.81 (95% CI, 0.553 to 1.183). The meaningful change threshold was 10 points.

PROMIS 7a

Fatigue was assessed using the patient-reported PROMIS 7a scale, which assessed the experience and impact of fatigue over the previous 7 days. Table 38 shows the mean scores at baseline and on day 1 of subsequent treatment cycles, the LS mean change from baseline, and the LS mean difference between venetoclax plus azacitidine and placebo plus azacitidine, for PROMIS 7a. Figure 5 shows the time to deterioration on the EORTC QLQ-C30 GHS/QoL. Values run from 0 to 100, with higher values indicating greater fatigue. At baseline, compliance in completing the PROMIS 7a was 93.6% and 92.3% in the venetoclax plus azacitidine and the placebo plus azacitidine groups, respectively. In subsequent cycles (up to cycle 13), compliance in individual treatment groups ranged from 72.4% to 84.9%.

Figure 5: Time to Deterioration on the PROMIS Fatigue 7a, Longitudinal Analysis Population

AZA = azacitidine; CI = confidence interval; GHS/QoL = global health status quality of life scale; HRQoL = health-related quality of life; LAP = longitudinal analysis population; EORTC QLQ-30C = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; PBO = placebo; PRO = patient-reported outcome; PROMIS Fatigue 7a = Patient-Reported Outcomes Measurement Information System Short Form v1.0–Fatigue 7a; VEN = venetoclax.

Source: Patient-reported outcome report.¹

The mean PROMIS 7a score at baseline was similar among patients in the venetoclax plus azacitidine group and the placebo plus azacitidine group: 53.86 (n = 264) and 54.97 (n = 132), respectively. There was a greater change from baseline in PROMIS 7a scores among patients in the venetoclax plus azacitidine group versus the placebo plus azacitidine group on day 1 of cycles 5, 7, 9, 11, and 13, but no clinically meaningful differences in mean change from baseline between treatment groups. This end point was part of the statistical testing hierarchy, but testing failed for an end point before this. Thus, conclusions cannot be drawn based on the results of this outcome without risk of increased type I error.

The median time to deterioration for patients randomized to venetoclax plus azacitidine was 9.3 months compared with 8.6 months for patients randomized to placebo plus azacitidine. The adjusted HR was 0.72 (95% CI, 0.509 to 1.011). The meaningful change threshold was 5 points.

EORTC QLQ-C30 Symptom Scales

The mean baseline EORTC QLQ-C30 fatigue subscale score for venetoclax plus azacitidine was 47.67 (n = 262) and for placebo plus azacitidine it was 49.83 (n = 130). Change from baseline showed improvement in subsequent cycles for both treatments, but the observed differences in LS means between the treatments were small at each time point. At baseline, compliance in responding to the questionnaire was 92.2% and 90.9% in the venetoclax plus azacitidine and the placebo plus azacitidine groups, respectively. In subsequent cycles (up to cycle 13), compliance in individual treatment groups ranged from 72.4% to 82.9%.

The mean baseline EORTC QLQ-30 nausea and vomiting subscale score for venetoclax plus azacitidine was 7.95 (n = 262) and for placebo plus azacitidine it was 7.63 (n = 130). Change from baseline showed improvement in subsequent cycles for both treatments, but the observed differences in LS means between the treatments were small at each time point.

The mean baseline EORTC QLQ-30 pain subscale score for venetoclax plus azacitidine was 21.95 (n = 262) and for placebo plus azacitidine it was 25.90 (n = 130). There was a numerically greater improvement for venetoclax plus azacitidine at most time points, but the observed differences in LS means between the treatments were small at each time point.

The mean baseline EORTC QLQ-30 appetite loss subscale score for venetoclax plus azacitidine was 31.42 (n = 262) and for placebo plus azacitidine it was 30.77 (n = 130). There was a greater improvement for venetoclax plus azacitidine from cycle 5 to cycle 13, but the observed differences in LS means between the treatments were small at each time point.

The mean baseline EORTC QLQ-30 constipation subscale score for venetoclax plus azacitidine was 20.23 (n = 262) and for placebo plus azacitidine it was 15.8 (n = 130). There was a greater improvement in subsequent cycles for venetoclax plus azacitidine, but the observed differences in LS means between the treatments were small at each time point.

The mean baseline EORTC QLQ-30 diarrhea subscale score for venetoclax plus azacitidine was 10.56 (n = 262) and for placebo plus azacitidine it was 9.74 (n = 130). There was improvement in subsequent cycles for venetoclax plus azacitidine for both treatments, but the observed differences in LS means between the treatments were small at each time point.

EuroQol Visual Analogue Scale

Table 38 shows the mean scores at baseline and day 1 of subsequent treatment cycles, the LS mean change from baseline, and the LS mean difference between venetoclax plus azacitidine and placebo plus azacitidine, for the EQ VAS. Figure 6 shows the time to deterioration on the EQ VAS.

The mean baseline EQ-5D-5L index score for venetoclax plus azacitidine was 0.76 (n = 260) and for placebo plus azacitidine it was 0.74 (n = 130). There was improvement (increase) in subsequent cycles for venetoclax plus azacitidine for both treatments, but the observed differences in LS means between the treatments were small at each time point. The mean baseline EQ VAS for venetoclax plus azacitidine was 60.29 (n = 260) and for placebo plus azacitidine it was 64.27 (n = 130). There was improvement (increase) in subsequent cycles for venetoclax plus azacitidine for both treatments, but the observed differences in LS means between the treatments were small at each time point.

The median time to deterioration for patients randomized to venetoclax plus azacitidine was 10.7 months, compared with 3.9 months for patients randomized to placebo plus azacitidine. The adjusted HR was 0.55 (95% CI, 0.394 to 0.768). The meaningful change threshold was 7 points.

Figure 6: Time to Deterioration on the EQ VAS, Longitudinal Analysis Population

AZA = azacitidine; LAP = longitudinal analysis population; EQ-5D = EuroQol 5-Dimensions questionnaire; EQ VAS = EuroQol Visual Analogue Scale; PBO = placebo; PRO = patient-reported outcome; VEN = venetoclax.

Source: PRO report.¹

Harms

Only those harms identified in the review protocol are reported subsequently. Table 14 shows an overall summary of the treatment-emergent adverse events, most common treatment-emergent adverse events, grade 3 or greater treatment-emergent adverse events, SAEs, adverse events leading to death, adverse events leading to discontinuation of venetoclax or placebo and azacitidine, and notable harms (January 4, 2020). Individual adverse events are in descending order of frequency for patients receiving venetoclax plus azacitidine.

All patients in both groups experienced at least 1 adverse event, and almost all experienced at least 1 grade 3 or greater adverse event. Compared with patients who received placebo plus azacitidine, a greater proportion of patients who received venetoclax plus azacitidine experienced 1 or more SAEs, 1 or more adverse events leading discontinuation or dose interruption for venetoclax or placebo or azacitidine, or 1 or more adverse events leading to death. Of the patients who received venetoclax plus azacitidine, 279 (98.6%) had grade 3 or greater adverse events, 235 (83.0%) had SAEs, 69 (24.4%) had adverse events leading to discontinuation of venetoclax or placebo, 68 (24.0%) had adverse events leading to discontinuation of azacitidine, and 64 (22.6%) had adverse events leading to death. Of the patients who received placebo plus azacitidine, 139 (96.5%) had grade 3 or greater adverse events, 105 (72.9%) had SAEs, 29 (20.1%) had adverse events leading to discontinuation of venetoclax or placebo, 29 (20.1%) had adverse events leading to discontinuation of azacitidine, and 29 (20.1%) had adverse events leading to death.

Treatment-Emergent Adverse Events

Table 14 shows treatment-emergent adverse events that were reported in 10% or more of patients and treatment-emergent adverse events reported in 5% or more of patients in the safety population at the time of IA2 (January 4, 2020). The most common adverse events were thrombocytopenia, nausea, constipation, neutropenia, febrile neutropenia, diarrhea, vomiting, hypokalemia, and anemia. The most common grade 3 or greater treatment-emergent adverse events were thrombocytopenia, neutropenia, febrile neutropenia, anemia, leukopenia, pneumonia, and hypokalemia.

The treatment-emergent adverse events reported with a frequency of 5% or greater in patients receiving venetoclax plus azacitidine compared with those receiving placebo plus azacitidine were thrombocytopenia, neutropenia, febrile neutropenia, anemia, leukopenia, nausea, diarrhea, vomiting, stomatitis, hemorrhoids, peripheral edema, asthenia; decreased appetite, arthralgia; dizziness, syncope, presyncope, vertigo; dyspnea; pruritus; and rash maculopapular.

Serious Adverse Events

Table 14 shows SAEs reported in 2% or more patients in the safety population at IA2 (January 4, 2020). The most common SAEs were febrile neutropenia, pneumonia, sepsis, anemia, and neutropenia. Febrile neutropenia, anemia, neutropenia, and thrombocytopenia were more frequent in patients who received venetoclax plus azacitidine, and pneumonia and sepsis were more frequent in patients who received placebo plus azacitidine.

Mortality

Table 14 shows adverse events leading to death for the safety population. A total of 64 (22.6%) patients who received venetoclax plus azacitidine and 29 (20.1%) who received placebo plus azacitidine had adverse events leading to death. The most frequent adverse events leading to death were pneumonia, sepsis, cardiac arrest, and an event recorded as death. Pneumonia, death, and cardiac arrest occurred more frequently in patients who received venetoclax plus azacitidine.

Treatment Discontinuations Due to Adverse Events

Table 14 shows adverse events leading to discontinuation of placebo or venetoclax for 2 or more patients in either group in the safety population by adverse event and system organ class.

The most frequent adverse events leading to venetoclax or placebo treatment discontinuation in the venetoclax plus azacitidine group were acute kidney injury, atrial fibrillation, febrile neutropenia, neutropenia, pneumonia, sepsis, thrombocytopenia, malignant neoplasm progression, and respiratory failure. In patients who received venetoclax plus azacitidine, atrial fibrillation, febrile neutropenia, cardiac failure, Klebsiella infection, and septic shock were more frequent, and in patients who received placebo plus azacitidine, pneumonia, sepsis, thrombocytopenia, and malignant neoplasm progression were more frequent.

The most frequent adverse events leading to azacitidine treatment discontinuation in the venetoclax plus azacitidine group were acute kidney injury, atrial fibrillation, neutropenia, pneumonia, sepsis, febrile neutropenia, and malignant neoplasm progression. In patients who received venetoclax plus azacitidine, atrial fibrillation, neutropenia, febrile neutropenia, cardiac failure, Klebsiella infection, and septic shock were more frequent, and in patients who received placebo plus azacitidine, pneumonia, sepsis, malignant neoplasm progression, and thrombocytopenia were more frequent.

Notable Harms

The notable harms identified for the protocol were neutropenia, febrile neutropenia, infections, tumour lysis syndrome, hemorrhage, and secondary malignancies. The search for adverse events of neutropenia included terms for neutropenia, neutrophil count decreased, febrile neutropenia, agranulocytosis, neutropenic infection, and neutropenic sepsis. Tumour lysis syndrome was a known risk and all patients received prophylaxis with oral and/or IV hydration and uric acid reducer, and were admitted for monitoring during ramp-up of venetoclax. Patients with a white blood cell count greater than 25 × 10⁹/L required cytoreduction before treatment.

Neutropenia, febrile neutropenia, infections and infestations, and secondary primary malignancies all occurred in a greater proportion of patients who received venetoclax plus azacitidine than in patients who received placebo plus azacitidine. Hemorrhage and tumour lysis syndrome occurred in similar proportions, and the proportion of patients with tumour lysis syndrome was low (≤ 2.5%).

The most common secondary malignancies were basal cell carcinoma and squamous cell carcinoma of the skin: 3 (1.1%) and 2 (0.7%) instances, respectively, in patients receiving venetoclax plus azacitidine, and none in patients receiving placebo plus azacitidine. Other malignancies reported for patients who received venetoclax plus azacitidine were gastric adenocarcinoma (recurrence), adenocarcinoma of colon, chloroma (manifestation of AML), erythroleukemia (AML), neuroendocrine carcinoma of the skin, and plasma cell myeloma. Other malignancies reported for patients receiving placebo plus azacitidine were malignant melanoma and renal cancer.

Critical Appraisal

Internal Validity

Randomization was conducted by an independent statistician, allocation was through an interactive voice recognition system (IVRS), and the study was double blinded, with identical-appearing oral venetoclax and placebo. Venetoclax does not appear to have any adverse events that are so specific as to unblind patients or physicians; therefore, the blinding was likely to remain intact. Blinding was not assessed.

There was no clinically meaningful imbalance in the baseline characteristics that might favour 1 group or the other, and there was minimal loss to follow-up (1.7% for venetoclax plus azacitidine and 1.4% for placebo plus azacitidine). The number of patients who withdrew consent after randomization was small (approximately 2%).

The duration of exposure and number of treatment cycles received was longer in the venetoclax plus azacitidine group, which is probably reflective of the difference in survival: patients receiving venetoclax plus azacitidine survived longer and received more cycles. A similar proportion of patients discontinued treatment due to adverse events. Procedures for assessing compliance with treatment (i.e., counting of returned doses) were described in the protocol but the results were not reported.

OS is a standard outcome in oncology drug investigation, with robust methods for ascertainment. Collection was likely to be complete and the timing of events was likely to be accurately determined. Standard methods for survival analysis were used, with surviving patients censored at the date they were known to be alive on or before the cut-off date. There was minimal loss to follow-up or withdrawal and good balance at baseline, so censoring is unlikely to be related to prognosis. The prognosis of the patients recruited is unlikely to have changed with time, as there were no changes to the inclusion and exclusion criteria that would be likely to affect prognosis, and recruitment took place over a relatively short time period.

EFS is a composite end point consisting of death from any cause, confirmed morphologic relapse from CR + CRi, confirmed disease progression, and treatment failure. Treatment failure was defined as failure to reach CR, CRi, or MLFS after at least 6 cycles. No protocol-specific support for the validity of the end point was offered in the protocol or statistical analysis plan. EFS is an accepted end point in the development of treatments for leukemia,²⁵ although empirical data show inconsistent correlation between EFS and OS.³⁵ However, it provides a more direct measurement of the ability of the treatment to achieve a response and the durability of the response achieved than OS, since EFS is affected by trial treatment alone, while OS is affected by trial treatment, post-trial treatment, and supportive or palliative care.³⁵ A time-to-event analysis of all individual end points making up the composite was not reported, making it difficult to fully assess for violations of the assumptions underlying the composite end points (i.e., the events were of equal importance to patients, occur with similar frequency, and have a similar sensitivity to the treatment). Death was reported for the greatest proportion of patients (54.5% and 61.5% for venetoclax plus azacitidine and placebo plus azacitidine, respectively), followed by confirmed morphologic relapse or confirmed disease progression (43.5% and 28.7% for venetoclax plus azacitidine and placebo plus azacitidine, respectively), and then by treatment failure (2.1% and 9.8% for venetoclax plus azacitidine and placebo plus azacitidine, respectively). The proportion of patients with each end point was reported, and the distribution for each treatment was consistent with the observed results for survival and CR + CRi, which were higher in the venetoclax plus azacitidine group than in the placebo plus azacitidine group. Results for individual analyses of OS, duration of response, and CR + CRi show similar directions of effect, but this does not adjust for competing events. Standard methods for survival analysis were used, with surviving patients censored at the date they were known to be alive on or before the cut-off date. There was minimal loss to follow-up or to withdrawal and a good balance at baseline, so censoring is unlikely to be related to prognosis. The prognosis of the patients recruited is unlikely to have changed with time, as there were no changes to the inclusion and exclusion criteria that would be likely to affect prognosis, and recruitment took place over a relatively short time period.

Composite complete remission (CR + CRi) and CR were investigator-assessed based on laboratory and clinical findings, with independent review. No protocol-specific support for the validity of the end point was offered in the protocol or statistical analysis plan. CR + CRi is an accepted end point in the development of treatments for leukemia,²⁵ although empirical data suggest the strength of the correlation between CR + CRi and OS may be population- and treatment-dependent.³⁵ The results were cross tabulated, and the differences minimal between investigator and independent review results. Treatment effect was not calculated for the end point and its 2 components, and statistical testing for CR + CRi and CR was performed at different interim analyses. CR and CRi both reflect bone marrow and peripheral blood improvement, with different thresholds, and the direction of effect was the same for both. Randomized patients without a post-baseline disease assessment were considered nonresponders. This is a conservative assumption, biasing the individual estimates of response downward, but accounts for a competing risk of death in an aged population.

Transfusion-independence rate was a pre-specified end point that was included in the sequential testing strategy. It is not clear how data on transfusion were collected, or whether these data might be susceptible to survivor bias, i.e., whether patients had to survive until the next transfusion visit for the previous visit to be captured. This risks undercounting transfusions in seriously ill patients. Patients who did not receive the study drug were considered transfusion-dependent, a conservative assumption affecting only a small number of patients.

Overall QoL was measured using the EOTRC QLQ-C30 GHS/QoL scale, cancer-related fatigue was measured using the PROMIS 7a, fatigue and other symptoms of interest were measured using EORTC QLQ-30C subscales, and health utility was measured using the EQ-5D-5L. The EOTRC QLQ-C30 GHS/QoL scale had previously been validated in mixed groups of cancer patients and an MID established. The PROMIS 7a has been validated in published studies of chronic illness, and the sponsor reported steps to validate the measure using the data from the VIALE-C study of venetoclax plus LDAC. Mean change from baseline was calculated using available data without imputation, and the level of compliance with the tool was not reported. Compliance post baseline was around 80%, meaning about 20% of available patients were not represented, and attrition due to death and disease progression was pronounced, meaning that later time points in particular represent a small survivor subgroup.

There were 2 preplanned interim analyses with preplanned stopping boundaries, with recommendations to stop or proceed made by an independent monitoring committee. Multiplicity due to interim analyses and the testing of multiple end points was controlled by a preplanned alpha-spending strategy and pre-specified hierarchy of testing with gatekeeping. Testing boundaries for selected end points representing proportions were adjusted by an information fraction to accommodate incomplete accrual at the interim analysis. Planned methods were reported on and adhered to.

Protocol violations included violations of inclusion and exclusion criteria, where patients were included although their bone marrow blast count did not meet the threshold at the time of testing for the study or the data were missing. A review provided evidence that they had previously met the criteria. Protocol violations in dosing included the use of strong and moderate CYP3A inhibitors without modifying the dose of venetoclax or placebo. These cases were reviewed for safety concerns. No violations were reported that substantially affected the internal validity of the study.

Subgroups of interest were pre-specified. Age (18 to < 75 years and ≥ 75 years) and cytogenetic risk (intermediate, poor) were used as stratification variables; other subgroups were not stratified. The aseline balance of all covariates was not reported for subgroups. Some subgroups were small, which was reflected in the high uncertainty of treatment estimates. This particularly affected estimates of CR + CRi, where both the number of specific events and total number of events could be small. Subgroup analyses of patients with IHD1, IHD2, and FLT3 mutations were included in the sequential testing strategy. Other subgroup analyses, including those on patients negative for mutations, were not adjusted for multiplicity.

There were 7 protocol amendments with corresponding changes in the statistical analysis plan. An early change to stratification factors meant that 2 patients randomized under the original protocol were not included in the efficacy analysis. Minor changes were made to inclusion criteria for safety reasons, there were clarifications on the dosing of concomitant medications, and the sample size was adjusted to allow for longer follow-up. Changes were made to the definition of EFS to align it with VIALE-C and to CR and CRi. All changes were made before the stopping and unblinding of the study following IA2 and are unlikely to have affected the internal validity of the study.

External Validity

The inclusion criteria for VIALE-A assumed that patients aged 75 years and older would not be eligible for standard induction chemotherapy, and age was the most common reason given for randomized patients being considered ineligible for standard induction chemotherapy. In Canadian practice, there is no consistency in defining an upper age limit for intensive chemotherapy. Chemotherapy may be considered for patients with treatment-naive AML aged 75 years and older, especially those with good or intermediate risk cytogenetics. The myelosuppressive nature of venetoclax plus azacitidine means it may not be suitable for frail patients or those who cannot travel for frequent lab visits, regardless of age.

Patients with CNS involvement were excluded, but clinical experts indicated that patients with CNS involvement might benefit from venetoclax plus azacitidine with concomitant intrathecal therapy. Patients with secondary AML (arising from prior myeloproliferative neoplasm including myelofibrosis, essential thrombocythemia, and polycythemia vera) were excluded from enrolling onto the study. However, there is data that shows activity of venetoclax plus azacitidine in this group of patients.^36,37 Patients with isolated granulocytic sarcoma were not included in the study.

Venetoclax dosing in the trial was aligned with the Health Canada–approved dosing. Up-titration and monitoring would be expected to be the same in clinical practice. Dosing of azacitidine was also aligned with Health Canada–indicated dosing; however, the current Health Canada approval for azacitidine is for patients with AML with less than 30% blasts. In practice, experts and clinician groups noted that jurisdictions are funding azacitidine and centres are already using it in patients with a blast count of 30% or greater. There is no maximum blast restriction in the Health Canada–approved indication for venetoclax plus azacitidine for patients with AML who are aged 75 years and older or ineligible for standard induction therapy, and the indication aligns with the submission. In the study, azacitidine was dosed for 7 consecutive days while, in practice, alternative dosing regimens are used to reduce the hematological toxicity, e.g., 5 to 2-2 and 6 consecutive days.

The outcome measures were relevant to patients and clinicians. They captured clinically important end points such as OS, clinically important surrogate end points of response and remission, overall QoL, and factors identified as influencing QoL (e.g., transfusion dependence). Disease remission, avoidance of relapse, symptoms, QoL, independence from transfusion and avoidance of hospitalization, and OS were all identified as important to patients and physicians. In clinical practice, strict responder definitions may not capture responding patients; patients may not reach strict response categories but may still derive clinical benefit.

The settings for the study were predominately urban hospitals and clinics. It therefore does not necessarily address the rural or remote Canadian context, where patients would not have access to frequent laboratory testing to monitor the ramp-up of venetoclax and cytopenias, nor access to outpatient or inpatient treatment for side effects and complications. Patients would be required to travel for treatment or to receive an alternative.

Duration of follow-up was around 20 months, with a median OS for venetoclax plus azacitidine of 14.7 months. Five-year follow-up is standard in oncology trials.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

An ITC was required because of a lack of studies directly comparing venetoclax plus azacitidine and venetoclax plus LDAC with other treatments currently in use in the Canadian setting.

Search Methods

A focused literature search for NMAs dealing with Venclexta (venetoclax) and AML was run in MEDLINE All (1946–) on February 11, 2021. No limits were applied.

Description of Indirect Treatment Comparison

One report that included ITCs was supplied by the sponsor. It included a systematic review with an NMA comparing venetoclax plus azacitidine and venetoclax plus LDAC with azacitidine, LDAC and BSC, and 2 propensity-score analyses comparing venetoclax plus azacitidine with LDAC (2-way comparison), and venetoclax plus azacitidine with azacitidine with LDAC (3-way comparison).

Table 15 shows the study selection criteria and key aspects of the methods for the systematic review. The patient population of interest included treatment-naive adult patients with AML who were ineligible for intensive chemotherapy, but the search allowed flexible wording to ensure retrieval of studies. Treatment naive was considered interchangeable with “previously untreated” or “newly diagnosed,” and “ineligible for chemotherapy” included patients described as old or elderly, unfit for intensive chemotherapy, unfit for standard chemotherapy, or unfit for high-dose chemotherapy. The initial search for articles included a broader set of comparators and included controlled clinical trials as a study design. More restricted selection criteria that were developed for a planned EUnetHTA submission were applied at the full-text review stage; the table reflects these criteria. The reasons for selection of comparators were not given, but the overall declared intention was to select high-quality studies that might enable ITCs.

Methods of the ITC

Objectives

The objective of this study was to compare the efficacy of venetoclax combination therapies with alternative treatments in treatment-naive patients with AML who were ineligible for intensive chemotherapy, including:

• Objective 1: Comparison of venetoclax plus azacitidine and venetoclax plus LDAC with azacitidine, LDAC, and BSC using NMA.

• Objective 2: Comparison of venetoclax plus azacitidine versus LDAC using propensity score–weighting analysis.

• Objective 3: Comparison of venetoclax plus azacitidine versus azacitidine versus LDAC using 3-way propensity score–weighting analysis.

Study Selection Methods

To be included in the NMAs, trials retrieved by the systematic review had to meet the following criteria:

• Study design: phase III RCTs

• Population: Treatment-naive adult patients with AML who were ineligible for intensive chemotherapy

• Interventions: Venetoclax plus azacitidine, venetoclax plus LDAC, LDAC, azacitidine, and BSC (including blood transfusion, etoposide, mercaptopurine, and hydroxyurea)

• Outcomes of interest: OS, EFS, CR, CRi, CR + CRi

The decision to restrict selection to phase III RCTs for reasons of quality led to the exclusion of trials containing glasdegib, as there was no phase III trial connected to the network containing venetoclax plus azacitidine and venetoclax plus LDAC.

ITC Analysis Methods

Three analyses were conducted: 1 NMA and 2 propensity score–weighted comparisons.

The NMA compared venetoclax plus azacitidine and venetoclax plus LDAC with comparators for the available end points of OS and CR + CRi. The feasibility of pooling to create a network for analysis was pre-assessed on the basis of study and patient characteristics. The main analysis excluded patients from the VIALE-C LDAC group who would not have been eligible to enter VIALE-A because they had previously been treated with an HMA or had good cytogenetic risk. For OS, the proportional hazards assumption was assessed using log-log cumulative hazard plots, which led to the decision to model OS using proportional hazards.

The model was a Bayesian mixed-treatment comparison using a generalized linear model framework, with OS modelled using the identity link and dichotomous outcomes modelled using the logit link. Due to limited data, only fixed-effects models were estimated. Prior distributions were non-informative and selected according to a process that was not detailed. Posterior probabilities were modelled using Markov chain Monte Carlo methods, with 50,000 iterations on 3 chains and a burn-in period of 50,000 iterations. Convergence was assessed using trace and density plots and Gelman-Rubin plots and diagnostics. Selection between models was made by the difference information criterion (DIC). The chosen definition of a meaningful difference in DIC was not given.

Two propensity score–weighting analyses were conducted. No specific rationale was provided for these additional analyses. The comparisons were:

• Two-way propensity-score weighting of individual patient data was used to compare to compare venetoclax plus azacitidine with LDAC, using the venetoclax plus azacitidine group from VIALE-A and the LDAC group from VIALE-C. Data were available for outcomes of OS, EFS, and CR + CRi. Individual patient data were available for both trials, with a data cut-off for VIALE-A of January 4, 2020 and for VIALE-C of August 15, 2019.

• Three-way propensity-score weighting of individual patient data was used to compare venetoclax plus azacitidine with azacitidine and with LDAC, using the venetoclax plus azacitidine and azacitidine groups from VIALE-A (which were randomized) and the LDAC group from VIALE-C. Data were available for outcomes of OS, EFS, and CR + CRi. Data cut-offs were the same as stated previously.

For both analyses, the propensity score for treatment was calculated for each patient using a logistic regression model with treatment with venetoclax plus azacitidine versus LDAC as the outcome and the baseline demographic and clinical characteristics as covariates. Analyses of OS, EFS, and CR + CRi were then conducted using data weighted by the inverse of the probability score. The covariates were:

• Demographic characteristics: Age (< 75 years, ≥ 75 years), sex (male, female), race (White, non-White).

• Clinical characteristics: AML type (primary and secondary), AML with myelodysplasia-related changes (yes, no), prior MDS (yes, no), bone marrow blasts (< 30%, ≥ 30%), cytogenetic risk (poor, intermediate), and ECOG PS (< 2, ≥ 2). Patients with missing values for any of these were excluded from the analysis.

For comparability, the propensity score–weighting analyses excluded patients from the VIALE-C LDAC group who would not have been eligible to enter VIALE-A: those who had been previously treated with an HMA and those who had good cytogenetic risk. For the 2-way propensity-score analysis, a subgroup analysis was conducted that was restricted to those patients in the main analysis who had greater than 30% bone marrow blasts at the baseline assessment, with a corresponding sensitivity analysis of all patients who had 30% bone marrow blasts. For the 3-way propensity-score analysis, a planned subgroup analysis of patients with 20% to 30% bone marrow blasts was not conducted, as the number of patients meeting these criteria in the LDAC arm was small.

Standard mean differences, t-tests (continuous variables), and chi-square tests (categorical variables) were used to assess balance before weighting, and weighted standard mean differences, weighted t-tests, and weighted chi-square tests were used to assess balance after weighting. An effective sample size was calculated and distributions of weights inspected to identify potential sensitivity to extreme weights. Time-to-event comparisons (OS and EFS) were made using weighted Cox proportional hazards models. Standard errors, 95% CIs, and P values were based on robust estimates of variances accounting for variability in propensity-score weights.

Results of ITC

Summary of Included Studies

Following removal of duplicates, 7,319 records were screened by title and abstract; of these, 225 were screened in full text. With the addition of the VIALE-A and VIALE-C study reports, the final selection was 7 RCTs with at least 2 arms of interest.

With the additional restriction of the comparators for the NMA inclusion criteria, removing decitabine from the comparators, 4 trials were included in the NMA: VIALE-A, VIALE-C, AZA-001, and AZA-AML-001. Table 16 shows a summary of the study characteristics for these 4 trials.

Table 17 shows a summary of patient baseline characteristics for the 4 studies included in the NMA. Only the arms used in the NMA are included. The table is in 2 panels, the first showing demographic and clinical characteristics and the second showing the cytogenetic and mutation data. Table 19 shows an assessment of heterogeneity based on the study and patient characteristics. The most important source of heterogeneity was in indicators of disease severity, bone marrow blast counts, proportion of patients with poor cytogenetic risk, and baseline ECOG PS.

Table 20 shows the results of the risk-of-bias assessment for the 4 trials included in the ITC. The quality assessment questions appear in Table 15. Risk of bias was low for all trials for treatment randomization, allocation concealment, and baseline balance. Trials AZA-001 and AZA-AML-001 were open-label, so the risk of bias was high, whereas VIALE-A and VIALE-C were double-blind, with a low risk of bias. AZA-001 was at high risk of bias for imbalance in drop-outs, as more patients appear to have dropped out of the conventional care arm, and there was selective reporting, as overall adverse events were not available. All were at unclear risk of bias for the inclusion of intention-to-treat analyses. Where a reason was given, the concern was with lack of detail on methods for handling missing data.

Trial Networks

Figure 7 shows the network for the NMA for OS. Four trials reported this end point and were included in the NMA. The network was linear with a single branch and included 5 treatments. There were no closed loops. Azacitidine was the best-represented treatments with 3 trials contributing data, followed by LDAC with 2.

Figure 7: Network Diagram for the NMA for OS

BSC = best supportive care; LDAC = low-dose cytarabine; NMA = network meta-analysis; OS = overall survival.

Source: Indirect treatment comparison report.³⁹

Figure 8 shows the network for the NMA for CR + CRi. Three trials reported this end point and were included in the NMA; the fourth trial, AZA-001, did not report data on CRi. The network was linear with a single branch and included 5 treatments. There were no closed loops. Azacitidine and LDAC were the best-represented treatments, with 2 contributing trials each.

Figure 8: Network Diagram for the NMA for CR + CRi

BSC = best supportive care; CR = complete remission; CRi = complete remission with incomplete hematological recovery; LDAC = low-dose cytarabine; NMA = network meta-analysis.

Source: Indirect treatment comparison report.³⁹

Table 21 shows the data included in the NMAs for the 2 end points of OS and CR + CRi. In the 2 trials comparing azacitidine with BSC, the HRs for OS were 0.60 (95% CI, 0.38 to 0.95) and 0.48 (95% CI, 0.24 to 0.94) for AZA-001 and AZA-AML-001, respectively. In the 2 trials comparing azacitidine with LDAC, the HRs for OS were 0.37 (95% CI, 0.12 to 1.13) and 0.90 (95% CI, 0.70 to 1.16) for AZA-001 and AZA-AML-001, respectively.

Results

Results of the NMA

OS: Table 22 shows the results for the NMA for OS. Venetoclax plus azacitidine was favoured over comparators azacitidine (HR = 0.66; 95% CrI, 0.52 to 0.85), LDAC (HR = 0.57; 95% CrI, 0.40 to 0.81), and BSC (HR = 0.37; 95% CrI, 0.24 to 0.58), with no treatment favoured between venetoclax plus azacitidine, and venetoclax plus LDAC (HR = 0.81; 95% CrI, 0.50 to 1.31).

CR + CRi: Table 23 shows the results for the NMA for OS. Venetoclax plus azacitidine was favoured over comparators azacitidine (OR 5.05; 95% CrI, 3.30 to 7.87), LDAC (OR = 5.42; 95% CrI, 2.80 to 10.50), and BSC (OR = 61.55; 95% CrI, 8.23 to 1,881.53), with no treatment favoured between venetoclax plus azacitidine and venetoclax plus LDAC (OR = 0.86; 95% CrI, 0.30 to 0.35).

Results of Propensity-Score Analyses: Venetoclax Plus Azacitidine Versus Low-Dose Cytarabine

Table 24 shows the baseline characteristics for the comparison between venetoclax plus azacitidine and LDAC for the analysis of the whole population, before and after weighting. Patients in the LDAC group with favourable cytogenetic risk or prior HMA use were excluded. The largest baseline imbalances in terms of standardized mean difference were in ECOG PS, secondary AML, and race, all of which were reduced by adjustment. This weighting was used for all efficacy analyses for this comparison in this population.

OS, overall population: Table 25 shows the results of the weighted and unweighted comparisons for OS for the comparison of venetoclax plus azacitidine and LDAC. Figure 9 shows the survival curves for both unweighted and weighted comparisons. Venetoclax plus azacitidine was favoured over LDAC in both the unweighted (HR = 0.47; 95% CI, 0.33 to 0.67) and weighted comparisons (HR = 0.50; 95% CI, 0.35 to 0.73). Median OS in the weighted comparison was 14.69 months (95% CI, 12.12 to 19.25) for venetoclax plus azacitidine compared with 7.43 months (95% CI, 3.15 to 10.18) for LDAC.

Figure 9: OS for Venetoclax Plus Azacitidine Versus LDAC for the Overall Population, Before and After Weighting

AZA = azacitidine; LDAC = low-dose cytarabine; VEN = venetoclax.

Source: Indirect treatment comparison report.³⁹

Event-free survival, overall population: Table 26 shows the results of the weighted and unweighted comparison for EFS between venetoclax plus azacitidine and LDAC. Figure 10 shows the survival curves for both weighted and weighted comparisons. Venetoclax plus azacitidine was favoured over LDAC in both the unweighted (HR = 0.40; 95% CI, 0.28 to 0.56) and the weighted comparisons (HR = 0.40; 95% CI, 0.28 to 0.58). Median EFS in the weighted comparison was 9.79 months (95% CI, 8.41 to 11.99) for venetoclax plus azacitidine compared with 3.06 months (95% CI, 1.71 to 5.82) for LDAC.

Figure 10: EFS for Venetoclax Plus Azacitidine Versus LDAC for the Overall Population, Before and After Weighting

AZA = azacitidine; EFS = event-free survival; LDAC = low-dose cytarabine.

Source: Indirect treatment report.³⁹

CR + CRi, overall population: Table 27 shows the results for the comparison between venetoclax plus azacitidine and LDAC for CR, CRi, and CR + CRi before and after weighting. Venetoclax plus azacitidine was favoured over LDAC for CR + CRi in both the unweighted (OR = 10.32; 95% CI, 4.67 to 22.89) and weighted comparison (OR = 10.17; 95% CI, 4.55 to 22.73). After weighting, the proportion of patients with CR + CRi was 0.66 (95% CI, 0.61 to 0.72) for venetoclax plus azacitidine and 0.16 (95% CI, 0.08 to 0.29) for LDAC.

Sensitivity analysis, overall population: The sensitivity analysis included all patients enrolled in the LDAC arm in VIALE-C (n = 66), regardless of prior HMA use or good cytogenetic risk. The results for the comparison of venetoclax plus azacitidine versus LDAC were consistent with the results from the main analysis, with an after-weighting OS HR of 0.47 (95% CI, 0.34 to 0.66), an EFS HR of 0.38 (95% CI, 0.28 to 0.43), and an OR for CR + CRi of 10.52 (95% CI, 4.90 to 22.58).

OS, subgroup with baseline bone marrow blast count of 30% or greater: Table 28 shows the baseline characteristics for the comparison between venetoclax plus azacitidine and LDAC for the analysis of the subpopulation of patients with 30% or greater blasts, before and after weighting. Patients in VIALE-C with favourable cytogenetic risk or prior HMA use were excluded. The largest baseline imbalances in terms of standard mean difference were in ECOG PS, secondary AML, and race, all of which were reduced by adjustment. This weighting was used for all efficacy analyses for this comparison in this subgroup.

OS, subgroup with baseline bone marrow blast count of 30% or greater: Table 29 shows the results of the weighted and unweighted comparison for OS for the comparison of venetoclax plus azacitidine versus LDAC in patients with a baseline bone marrow blast count of 30% or greater. Venetoclax plus azacitidine was favoured over LDAC in both the unweighted (HR = 0.47; 95% CI, 0.33 to 0.67) and the weighted comparisons (HR = 0.47; 95% CI, 0.32 to 0.69). Median OS in the weighted comparison was 14.06 months (95% CI, 10.61 to 17.15) for venetoclax plus azacitidine compared with 3.61 months (95% CI, 3.12 to 10.18) for LDAC.