Reimbursement Review

Crovalimab (Piasky)

Sponsor: Hoffmann-La Roche Limited

Therapeutic area: Paroxysmal nocturnal hemoglobinuria

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Abbreviations

Executive Summary

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

Table 1: Background Information on Application Submitted for Review

NOC = Notice of Compliance; PNH = paroxysmal nocturnal hemoglobinuria; SC = subcutaneous.

Introduction

Paroxysmal nocturnal hemoglobinuria (PNH) is a rare, life-threatening, complement-mediated hematological disorder. PNH is the primary manifestation of chronic intravascular hemolysis, associated with bone marrow failure and thrombophilia.¹ It develops when hematopoietic cells acquire somatic mutations in the X-linked gene of PIGA.² Key clinical signs and symptoms of PNH include anemia, fatigue, thrombosis, esophageal spasms, male erectile dysfunction, hemoglobinuria, abdominal pain, and dyspnea.³ Patients with PNH may experience impaired health-related quality of life (HRQoL), and may struggle to complete normal everyday activities.^1,4 The mortality rate of patients with PNH who receive supportive care alone is approximately 20% to 35% within 5 years to 10 years of diagnosis.⁴ Chronic kidney disease, pulmonary hypertension, and venous or arterial thromboembolic events are life-threatening complications of PNH.⁴ Thromboembolic events were the leading cause of death in patients with PNH (40% to 67% of deaths with known cause).⁵ An allogeneic bone marrow transplant is considered a curative treatment for PNH.^6,7 However, a bone marrow transplant is not usually offered as a first-line treatment for classic PNH, owing to the risks of transplant-related morbidity and mortality, and is usually reserved for younger patients with PNH with concomitant severe aplastic anemia or other causes of bone marrow failure.^4,8,9 In Canada, patients with PNH receive C5 inhibitors as standard first-line therapy to reduce uncontrolled complement activation in the terminal complement cascade, and the primary C5 inhibitor therapies are ravulizumab and eculizumab, both administered through IV infusion.⁴

The objective of this Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of crovalimab 340 mg/2 mL for subcutaneous (SC) injection following 1 loading dose by IV infusion in the treatment of PNH in adults and adolescents aged 13 years and older with a body weight of at least 40 kg.

Perspectives of Patients, Clinicians, and Drug Programs

The information in this section is a summary of input provided by the patient and clinician groups that responded to the call by Canada’s Drug Agency (CDA-AMC) for input and from clinical experts consulted for the purpose of this review.

Patient Input

CDA-AMC received 1 joint input from the Canadian Association of PNH Patients and the Aplastic Anemia & Myelodysplasia Association of Canada (AAMAC). The patient groups gathered personal experiences from 1 patient living in Canada who had received crovalimab and provided information from the crovalimab clinical trial to highlight the trial’s findings and demonstrate the potential efficacy of crovalimab. The patient group input stated that PNH is a complex and multifaceted disorder that requires comprehensive management to address the various aspects of the disease and improve the quality of life (QoL) for those affected. The chronic nature of PNH means that patients must manage their condition over a lifetime, dealing with the physical, emotional, and financial burdens associated with the disease. The impact on QoL is profound, as patients must cope with the unpredictability of symptoms, the side effects of treatments, and the constant threat of serious complications. Regular monitoring and supportive care are essential to managing the disease and improving patient outcomes. According to the patient group input, 1 patient with PNH in Canada reported that the disease started with the patient feeling constantly exhausted, followed by dark urine in the morning and chest pain. The patient explained feeling terrified of the risk of blood clots and not knowing the effect of the disease on life, job, and family, along with accepting that this is a lifelong condition. Based on the patient group input, 1 significant challenge with current treatments is that they require regular clinic visits every 2 weeks to 8 weeks for IV infusions, which can be time-consuming and disruptive. The patient group stated that crovalimab has the potential to greatly improve patients’ QoL. It reduces the physical, emotional, and logistical burdens associated with traditional IV therapies, allowing patients to regain a sense of freedom and control over their lives. Fewer clinic visits and less travel for IV injections mean lower health care costs and less financial burden for patients. This more convenient and less invasive approach can significantly improve not only patients’ QoL but also their overall well-being and adherence to treatment, leading to better health outcomes.

Clinician Input

Input From Clinical Experts Consulted for This Review

The clinical experts consulted for this review noted that crovalimab may address an unmet need for some patients with a preference for an SC route of drug administration, particularly for those with poor venous access and who are opposed to the insertion of a port-a-cath (a tube that is inserted into a vein and connected to a port that is placed under the skin). The clinical experts noted that crovalimab may be a safer alternative for patients living in isolated geographic locations that preclude the guaranteed delivery of eculizumab, and some patients may be more comfortable with self-administration than having to travel to a clinic for IV therapy (which is associated with transportation costs and exposure to potential infections at a clinic). The clinical experts noted that where ravulizumab is funded, most patients with PNH would choose to start ravulizumab as the first-line therapy over eculizumab or crovalimab (of note, ravulizumab is not funded in some jurisdictions such as British Columbia). However, per clinical guidelines, eculizumab would be preferred for patients who are pregnant. Furthermore, the clinical experts pointed out that patients who prefer to be more independent in receiving therapy, have significant needle phobia, or find it difficult to visit a clinic due to geographic location may choose crovalimab as first-line therapy. The clinical experts noted that switching from ravulizumab or eculizumab to crovalimab could be uncommon as there is a risk of transient immune complex formation that only occurs in patients switching to or from other C5 inhibitors. The clinical experts further pointed out that in children, due to the high association with aplastic or hypoplastic anemia, management would be driven by treating aplastic or hypoplastic anemia. The clinical experts noted that the rare cases of isolated PNH would be treated along the same lines as for adults, except that PNH clone size is smaller in children and only clones greater than 10% would be eligible for C5 inhibition therapy. The clinical experts noted that patients who are fully motivated to self-administer the SC injections and those known to have genetic polymorphism that obviates efficacy with eculizumab or ravulizumab would be best suited for treatment with crovalimab. The clinical experts pointed out that the outcomes used to assess the treatment response would be the same as for eculizumab or ravulizumab, including improvements in hemoglobin levels, transfusion avoidance, and renal function, along with reductions in smooth muscle spasm, less fatigue, the normalization of lactate dehydrogenase (LDH), and the avoidance of breakthrough hemolysis (BTH). The clinical experts noted that frequent laboratory and patient assessments are required during the early stages of crovalimab, followed by standard assessments once the laboratory results stabilize. The clinical experts noted that if patients treated with crovalimab are not showing any clinical responses by 3 weeks to 4 weeks, a transition back to ravulizumab or eculizumab and then monitoring for the immune responses should be considered. The clinical experts pointed out that crovalimab should be discontinued when thrombosis occurs, or when there is ongoing LDH elevation or ongoing symptoms of PNH.

Clinician Group Input

CDA-AMC received 1 clinician group input from the Canadian PNH Network. Information for this submission was obtained via publicly available documents, congress abstracts, the published literature, and members of the Canadian PNH Network who were invited to contribute to the various sections. Three clinicians contributed to this submission. According to the clinician group input, the current standard of care for patients with hemolytic PNH is terminal complement inhibition with a C5 blockade; eculizumab and ravulizumab remain the only first-line therapies across Canada. The only curative treatment for PNH is an allogeneic hematopoietic stem cell transplant, which is reserved for patients with predominant or progressive bone marrow failure (e.g., aplastic anemia), which can coincide with, precede, or follow a diagnosis of PNH.

The Canadian PNH Network highlighted that crovalimab is expected to address important treatment goals unmet by current therapies, including offering a C5 inhibitory strategy that does not require IV access, enabling self-administration by patients or administration by caregivers, and providing an option for rare cases of resistance to eculizumab or ravulizumab due to a C5 polymorphism. Additionally, the clinician group expected that crovalimab would be another first-line option for patients with PNH, either switching from an IV C5 inhibitor or starting from being treatment-naive. Patients who favour the freedom and reduced treatment burden of SC administration would very likely select this therapy. On the other hand, patients with PNH least suitable would be those who are not accepting SC drug delivery, or those who develop clinically significant extravascular hemolysis, which necessitates proximal complement inhibition strategies.

The Canadian PNH Network stated that response to treatment focuses on LDH reduction, which reduces hemolysis and the risk of thrombosis, and may also improve hemoglobin levels and reduce transfusion dependence in patients with PNH. Clinical outcomes related to the response included decreased fatigue and transfusion requirements as well as improved QoL and overall survival. Efficacy outcomes would typically be assessed every 2 weeks to 4 weeks initially after starting a new therapy or switching, but follow-up would be required less often (e.g., every 3 months to 6 months) as a patient becomes established on the drug and does not show evidence of side effects or other concerns.

Based on the input, some of the factors that should be considered to discontinue the treatment include adverse events (AEs), type III immune complex reactions, poor compliance, and pregnancy. The clinician group added that treatment with crovalimab is most likely going to be done either entirely at the patient’s home or, in the case of loading doses, at a local infusion clinic.

Drug Program Input

Input was obtained from the drug programs that participate in the reimbursement review process. The following were identified as key factors that could potentially impact the implementation of a Canadian Drug Expert Committee recommendation for crovalimab:

• considerations for the initiation of therapy

• considerations for the prescribing of therapy.

The clinical experts consulted for this review provided advice on the potential implementation issues raised by the drug programs. Refer to Table 5 for further details.

Clinical Evidence

Systematic Review

Description of Studies

Two multicentre, phase III, randomized, open-label, active-controlled trials submitted by the sponsor were included, comparing crovalimab (maintenance dose is 680 mg [body weight ≥ 40 kg to < 100 kg] or 1,020 mg [≥ 100 kg] by SC injection once every 4 weeks) with eculizumab (maintenance dose is 900 mg by IV infusion once every 2 weeks) in adult patients with PNH who had not been previously treated with a complement inhibitor (the COMMODORE 2 study, randomized N = 204) and patients with documented treatment with complement inhibitors (the COMMODORE 1 study, randomized N = 89). Patients were randomized to receive either crovalimab (arm A) or eculizumab (arm B) in a 2:1 ratio in the COMMODORE 2 study, and in a 1:1 ratio in the COMMODORE 1 study. In addition, both trials enrolled a nonrandomized descriptive arm (arm C) where patients only received crovalimab (with 6 patients with PNH aged younger than 18 years in the COMMODORE 2 study and 38 patients on a complement inhibitor treatment, including 1 patient with PNH aged under 18 years, in the COMMODORE 1 study, at the clinical cut-off date of November 16, 2022). In both trials, the primary treatment period was 24 weeks for all arms. After 24 weeks, all patients had the opportunity to continue or switch to crovalimab in a 46-week safety follow-up period. In the COMMODORE 2 trial, transfusion avoidance and hemolysis control (measured by an LDH level ≤ 1.5 × upper limit of normal [ULN] at a central laboratory) were coprimary outcomes. Secondary outcomes were BTH, stabilized hemoglobin, and the Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-F) version 4 score. The mean percentage change in LDH levels, maintenance of a minimum hemoglobin level, and the number of pRBC units transfused, among others, were also reported as exploratory outcomes. HRQoL assessed with the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) version 3 tool and selected symptoms of the European Organisation for Research and Treatment of Cancer – Item Library 40 (EORTC IL-40) were evaluated. Patient preference for crovalimab or eculizumab, and harms were also reported. In the COMMODORE 1 trial, due to the introduction of ravulizumab to the treatment landscape and a reduced pool of patients treated with eculizumab over time, randomization was stopped in November 2022 per the amended protocol, which was earlier than initially planned in the study. As a result, the evaluation of safety became the primary objective and all efficacy objectives became exploratory end points. The specific safety and efficacy outcomes in the COMMODORE 1 trial were similar to those in the COMMODORE 2 trial.

At baseline, the median age of the randomized population in the crovalimab and eculizumab arms was 36 years (range, 18 years to 76 years) and 38 years (range, 17 years to 78 years), respectively, in the COMMODORE 2 study, and 42 years (range, 21 years to 81 years) and 49 years (range, 22 years to 85 years), respectively, in the COMMODORE 1 study. The proportions of females and males were similar between the 2 arms in both studies (43% to 49% were females and 51% to 57% were males in the COMMODORE 2 study, and 50% to 53% were females and 47% to 50% were males in the COMMODORE 1 study). Most patients were Asian (64% in the crovalimab arm and 74% in the eculizumab arm) or white (33% in the crovalimab arm versus 23% in the eculizumab arm), followed by other or unknown (3%) in the COMMODORE 2 trial. In the COMMODORE 1 trial, most patients were white (76% in the crovalimab arm versus 73% in the eculizumab arm), followed by Asian (20% in the crovalimab arm versus 16% in the eculizumab arm), and other or unknown (4% in the crovalimab arm versus 11% in the eculizumab arm). In the COMMODORE 2 trial, the baseline mean LDH level was 7.6 multiplied by ULN (standard deviation [SD] = 3.4 × ULN) and 7.8 multiplied by ULN (SD = 3.5 × ULN) in the crovalimab and eculizumab arms, respectively. In the COMMODORE 2 study, a total of 77% of patients in the crovalimab arm had received a pRBC transfusion within the 12 months before screening, with a mean of 6.5 (SD = 8.3) transfused pRBC units; 74% of patients in the eculizumab arm had received a pRBC transfusion with a mean of 6.6 (SD = 8.7) transfused pRBC units (half of the patients received > 0 to ≤ 6 transfused pRBC units in the prior 6 months in both arms). In the COMMODORE 1 trial, 23% of the patients in the crovalimab arm had received a mean of 1.6 (SD = 3.7) transfused pRBC units, and 25% of the patients in the eculizumab arm had received a mean of 2.3 (SD = 5.4) transfused pRBC units. In both studies, the median of the PNH clone size was smaller in the crovalimab arm than in the eculizumab arm for erythrocytes (25% in the crovalimab arm versus 45% in the eculizumab arm in the COMMODORE 2 study and 45% in the crovalimab arm versus 47% in the eculizumab arm in the COMMODORE 1 trial), for granulocytes (60% in the crovalimab arm versus 75% in the eculizumab arm in the COMMODORE 2 trial and 67% in the crovalimab arm versus 68% in the eculizumab arm in the COMMODORE 1 trial), and for monocytes (91% in the crovalimab arm versus 95% in the eculizumab arm in the COMMODORE 2 trial and 89% in the crovalimab arm versus 96% in the eculizumab arm in the COMMODORE 1 trial). The main PNH-relevant conditions in patient history were aplastic anemia (39% in the crovalimab arm versus 38% in the eculizumab arm in the COMMODORE 2 study and 33% in the crovalimab arm versus 36% in the eculizumab arm in the COMMODORE 1 study) and major vascular events (16% in the crovalimab arm versus 15% in the eculizumab arm in the COMMODORE 2 study and 22% in the crovalimab arm versus 23% in the eculizumab arm in the COMMODORE 1 study).

Efficacy Results

The key efficacy results from the COMMODORE 2 and COMMODORE 1 studies are summarized in Table 2 and Table 3, respectively, ordered by importance and with categories suggested by the clinical experts consulted for this review. In the COMMODORE 1 study, all efficacy outcomes are exploratory as randomization was stopped early, and there was not sufficient power for formal statistical noninferiority or superiority testing.

Hemolysis Outcomes

Hemolysis control: In the COMMODORE 2 study, crovalimab demonstrated noninferiority compared with eculizumab for hemolysis control as measured by a central LDH level of 1.5 or less multiplied by ULN (odds ratio [OR] = 1.02; 95% confidence interval [CI], 0.57 to 1.82; P value = ████) based on a predefined noninferiority margin (NIM) of 0.2 for the lower limit of the 95% CI. The mean proportion of patients with hemolysis control from week 5 through week 25 was 79.3% (95% CI, 72.9% to 84.5%) in the crovalimab arm and 79.0% (95% CI, 69.7% to 86.0%) in the eculizumab arm.

In the COMMODORE 1 study, the proportion of patients with a central LDH level of 1.5 or less multiplied by ULN at baseline was 93.2% for crovalimab and 95.2% for eculizumab. The mean proportion of patients who achieved hemolysis control (a central LDH level ≤ 1.5 × ULN) during the primary treatment period (i.e., from baseline through week 25) was 92.9% (95% CI, 86.6% to 96.4%) in the crovalimab arm and 93.7% (95% CI, 87.3% to 97.0%) in the eculizumab arm.

BTH: In the COMMODORE 2 study, crovalimab showed noninferiority efficacy compared with eculizumab in the proportion of patients with a BTH event from baseline through week 25 with the crovalimab arm at 10.4% (95% CI, 6.0% to 17.2%) versus the eculizumab arm at 14.5% (95% CI, 7.5% to 25.5%) with a weighted difference of −3.9% (95% CI, −14.8% to 5.3%; P value = ████) based on a predefined NIM of 20% for the upper limit of the 95% CI.

In the COMMODORE 1 study, the proportion of patients with BTH from baseline through week 25 was 10.3% (95% CI, 3.3% to 25.2%) in the crovalimab arm and 13.5% (95% CI, 5.1% to 29.6%) in the eculizumab arm.

Mean percentage change in LDH levels: In the COMMODORE 2 study, the mean percentage reduction in LDH levels from baseline to week 25 was −73.6% (95% CI, −78.95% to −68.2%) in the crovalimab arm and −64.1% (95% CI, −71.4% to −56.8%) in the eculizumab arm.

In the COMMODORE 1 study, the mean percentage change in the central LDH level from baseline to the average over week 21, week 23, and week 25 was 16.6% (95% CI, ███ ██ ████) in the crovalimab arm and 4.5% (95% CI, ████ ██ ████) in the eculizumab arm.

Hemoglobin Outcomes

Stabilized hemoglobin: In the COMMODORE 2 study, crovalimab showed noninferiority efficacy compared with eculizumab in the proportion of patients reaching hemoglobin stabilization (avoidance of a ≥ 2 g/dL decrease in the hemoglobin level from baseline, in the absence of transfusion) from baseline through week 25 with the crovalimab arm at 63.4% (95% CI, 54.6% to 71.5%) versus the eculizumab arm at 60.9% (95% CI, 48.4% to 72.2%) with a weighted difference of 2.2% (95% CI, −11.4% to 16.3%; P value = ████) based on a predefined NIM of −20% for the lower limit of the 95% CI.

In the COMMODORE 1 study, the proportion of patients with stabilized hemoglobin from baseline through week 25 was 59.0% (95% CI, 42.2% to 74.0%) in the crovalimab arm versus 70.3% (95% CI, 52.8% to 83.6%) in the eculizumab arm.

Patients Who Reached and Maintained a Minimum Hemoglobin Level (Hemoglobin ≥ 10 g/dL Without a Subsequent Decrease < 9 g/dL)

The proportion of patients who reached and maintained a minimum hemoglobin level from baseline through week 25 was █████ ████ ███ ████ ██ █████ in the crovalimab arm and █████ ████ ███ ████ ██ █████ in the eculizumab arm in the COMMODORE 2 study; in the COMMODORE 1 study, the proportion was █████ ████ ███ ████ ██ █████ in the crovalimab arm and █████ ████ ███ ████ ██ █████ in the eculizumab arm.

Transfusion Outcomes

Transfusion avoidance: In the COMMODORE 2 study, crovalimab demonstrated noninferiority compared with eculizumab for transfusion avoidance from baseline through week 25 with a weighted difference in the proportion of patients with transfusion avoidance of −2.8% (95% CI, −15.7% to 11.1%, P = ████), based on the predefined NIM of −20% for the lower limit of the 95% CI. In the crovalimab arm, 65.7% (95% CI, 56.9% to 73.5%) of patients were transfusion-free from baseline through week 25 compared with 68.1% (95% CI, 55.7% to 78.5%) of patients in the eculizumab arm.

In the COMMODORE 1 study, the proportion of patients who achieved transfusion avoidance from baseline through week 25 was 79.5% (95% CI, 63.1% to 90.1%) for the crovalimab arm and 78.4% (95% CI, 61.3% to 89.6%) for the eculizumab arm.

Red blood cell transfusions (packed red blood cell [pRBC] units transfused): In the COMMODORE 2 study, during the 24-week randomized primary treatment period, 45 (33.6%) patients in the crovalimab arm and 22 (31.9%) patients in the eculizumab arm had at least 1 transfusion. The mean number of units of pRBCs transfused was 2.3 units (95% CI, 1.3 units to 3.4 units) and 2.2 units (95% CI, 1.0 unit to 3.4 units) in the crovalimab and eculizumab arms, respectively.

In the COMMODORE 1 study, 8 (20.5%) patients in the crovalimab arm and 7 (18.9%) patients in the eculizumab arm had at least 1 transfusion, and the mean number of pRBC units transfused from baseline to week 25 in all randomized patients was 0.97 unit (95% CI, 0.2 unit to 1.7 units) in the crovalimab arm and 1.9 units (95% CI, 0.5 unit to 3.3 units) in the eculizumab arm.

Patient-Reported Outcomes

FACIT-F: In both the COMMODORE 2 and COMMODORE 1 studies, FACIT-F (range, 0 to 52, with a higher score indicative of less fatigue) was assessed in adult patients only (≥ 18 years).

In the COMMODORE 2 study, FACIT-F data were evaluable in 95.5% of adult patients in the crovalimab arm and 95.7% of adult patients in the eculizumab arm at each visit from baseline through week 25. The adjusted mean change (improvement) from baseline at week 25 in FACIT-F was 7.8 points (95% CI, 6.5 points to 9.1 points) in the crovalimab arm compared to 5.2 points (95% CI, 3.4 points to 6.9 points) in the eculizumab arm, with a between-group difference of 2.6 points (95% CI, 0.7 point to 4.6 points). By week 25, 58.6% of patients in the crovalimab arm and 54.5% of patients in the eculizumab arm had an improvement in fatigue severity of at least 5 points.

In the COMMODORE 1 study, FACIT-F data were evaluable in 86.4% of adult patients in the crovalimab arm and 76.2% of adult patients in the eculizumab arm at each visit from baseline through week 25. The adjusted mean change in FACIT-F scores from baseline to week 25 was 1.1 points (95% CI, −1.5 points to 3.7 points) in the crovalimab arm and −2.6 points (95% CI, −5.4 points to 0.1 point) in the eculizumab arm, with a between-group difference of 3.7 points (95% CI, 0.05 point to 7.4 points). The proportion of patients with a 5-point or greater improvement from baseline in the FACIT-F score was not reported for the COMMODORE 1 study.

EORTC QLQ-C30 global health status (GHS) and QoL: In the COMMODORE 2 study, the mean change in the EORTC QLQ-C30 GHS and QoL score from baseline to week 25 was 13.4 points (95% CI, 10.1 points to 16.7 points) for the crovalimab arm and 9.9 points (95% CI, 4.8 points to 14.9 points) for the eculizumab arm. Mean values at week 25 were similar to normative population values.

In the COMMODORE 1 study, the mean change in the EORTC QLQ-C30 GHS and QoL score from baseline to week 25 was 5.7 points (95% CI, −2.4 points to 13.8 points) in the crovalimab arm and −1.0 point (95% CI, −6.9 points to 4.9 points) in the eculizumab arm.

Harms Results

During the 24-week primary safety period, in the COMMODORE 2 study, there were similar proportions of patients in the 2 treatment arms who reported at least 1 AE, with 78% of patients in the crovalimab arm and 80% of patients in the eculizumab arm. A total of 10% of patients in the crovalimab arm and 13% of patients in the eculizumab arm reported at least 1 serious adverse event (SAE). The most common AEs were any infections (24% in the crovalimab arm and 36% in the eculizumab arm), infusion-related reactions (16% in the crovalimab arm versus 13% in the eculizumab arm), hypokalemia (11% in the crovalimab arm versus 13% in the eculizumab arm), hypersensitivity other than type III immune complex reactions (type III hypersensitivity reactions, ██ ███ ████), and injection-related reactions (5% in the crovalimab arm versus not applicable in the eculizumab arm). The most common SAEs were infections and infestations (3% in the crovalimab arm versus 7% in the eculizumab arm).

In the COMMODORE 1 study, a higher proportion of patients in the crovalimab arm compared with the eculizumab arm reported at least 1 AE (77% versus 67%, respectively) or at least 1 SAE (14% versus 2%, respectively). The most common AEs in the crovalimab arm compared with the eculizumab arm were any infections (41% versus 36%, respectively), infusion-related reactions (14% versus 0%, respectively), hypersensitivity other than type III immune complex reactions (██ ███), and injection-related reactions (7% versus not applicable, respectively). The most common SAEs were infections and infestations (7% in the crovalimab arm versus 2% in the eculizumab arm).

In both trials, no cases of infection with Neisseria meningitidis, including meningococcal meningitis, were reported in either arm.

One patient each from the crovalimab arm (0.7%) and the eculizumab arm (1.4%) in the COMMODORE 2 study and no patients in the COMMODORE 1 study experienced AEs leading to withdrawal from treatment with crovalimab or eculizumab.

In the COMMODORE 2 study, no adverse events of special interest (AESIs) of type III immune complex reactions related to drug-target-drug complexes (DTDCs) were reported in either arm during the primary treatment period as the patients were treatment-naive. In the COMMODORE 1 study, 7 (15.9%) patients in the crovalimab arm versus no patients in the eculizumab arm experienced type III immune complex reactions, because these AEs were expected only in patients who switched from eculizumab to crovalimab who are at risk of developing DTDC-associated type III immune complex reactions. In both trials, there were no AESIs reported for abnormal liver function tests or suspected transmission of an infectious drug by the study drug in either arm.

In the COMMODORE 2 study, death was reported in 2 (1.5%) patients in the crovalimab arm (1 patient died of respiratory tract hemorrhage and the other patient died of myocardial infarction) and 1 (1.4%) patient in the eculizumab arm (this patient died of ischemic stroke). In the COMMODORE 1 study, no deaths were reported during the primary safety period.

Critical Appraisal

Methods for randomization and allocation concealment were appropriate in both the COMMODORE 2 and COMMODORE 1 trials. In both trials, most baseline characteristics were similar between the randomized crovalimab and eculizumab arms except for numeric differences in a few of the characteristics. In the COMMODORE 2 study, they included race (64% Asian and 33% white in the crovalimab arm, and 74% Asian and 23% white in the eculizumab arm), the median PNH clone size for erythrocytes (25% in the crovalimab arm and 45% in the eculizumab arm) and for granulocytes (60% in the crovalimab arm versus 75% in the eculizumab arm), as well as the history of myelodysplastic syndrome (MDS) before enrolment (4% in the crovalimab arm versus 9% in the eculizumab arm); in the COMMODORE 1 trial, numeric differences in certain characteristics included the median PNH clone size for monocytes (89% in the crovalimab arm versus 96% in the eculizumab arm). Nevertheless, the clinical experts did not consider that the between-group imbalance in these characteristics would impact the efficacy and safety results in the studies. In the COMMODORE 2 study, the predefined NIMs for the efficacy outcomes were considered appropriate. In the COMMODORE 1 study, randomization was stopped early (in November 2022 per protocol amendment version 6) due to the introduction of ravulizumab to the treatment landscape, and a reduced pool of patients treated with eculizumab over time.¹⁰ The initially targeted sample size for the randomized arms of approximately 200 patients could not be reached, providing insufficient statistical power for efficacy analyses. The results of the exploratory efficacy analyses were reported descriptively, with no formal statistical noninferiority or superiority testing, limiting the interpretation and certainty of efficacy results of the COMMODORE 1 study.

The clinical experts consulted for this review noted that the maintenance dose of eculizumab (900 mg maintenance every 2 weeks for IV infusion) in both pivotal studies aligned with its product monograph.¹¹ In clinical practice, when BTH occurs repetitively, increasing the eculizumab dose may be considered, according to the clinical experts. For the COMMODORE 2 study, the lack of permitted dosage adjustments may have biased the efficacy results in favour of crovalimab relative to how eculizumab is dosed in clinical practice; however, the magnitude of this potential bias is unclear. In the pivotal trials, there was no randomized comparative evidence of crovalimab versus eculizumab at a higher dose than that recommended by Health Canada in patients who are C5 inhibitor–naive and in patients who switched from eculizumab.

Fatigue and HRQoL results were evaluable in most patients in the COMMODORE 2 study (96% of patients) and in 76% to 86% of patients in the COMMODORE 1 study. For these outcomes, the impact of missing outcome data in the COMMODORE 2 study is minimal; in the COMMODORE 1 study, the impact of missing outcome data (14% in the crovalimab arm and 24% in the eculizumab arm) is unclear. The open-label design of both trials may have impacted the reporting of subjective patient-reported outcomes (PROs), including the FACIT-F score, the EORTC QLQ-C30 GHS and QoL score, EORTC IL-40 symptoms, and patient treatment preference.

The interpretation of results of arm C in both trials is limited due to the lack of a comparator, the limited number of patients and events, and the descriptive summary of data. Similarly, the lack of a comparator arm does not allow for a conclusion to be drawn on the effect of crovalimab versus any comparator during the extension periods beyond week 25 in both the COMMODORE 2 and COMMODORE 1 trials.

The clinical experts noted that overall, the eligibility criteria of patients in both trials aligned with the diagnosis standard and treatment indication for PNH in clinical settings, and the demographic and disease characteristics (including LDH levels and history of transfusion) of the patients were mostly aligned with the patients seen in clinical practice in Canada. The clinical experts consulted for this review noted that the patients excluded from the COMMODORE 2 and COMMODORE 1 trials due to not meeting 1 of the patient inclusion criteria (granulocyte or monocyte clone size ≥ 10%) are typically asymptomatic and would less likely be considered for treatment. If these patients were showing other disease manifestations indicating a therapy, crovalimab could also be used with anticipated similar efficacy and safety results as those seen in the 2 trials as per the clinical experts.

The sample size of pediatric patients was small in the COMMODORE 2 study (n = 6) and in the COMMODORE 1 study (n = 1), and only descriptive results were available. The clinical experts expected that pediatric patients would have efficacy results similar to those in the main trial arms; however, there is a need for enhanced safety consideration regarding the risk of infections, including meningitis in pediatric patients. The recommended body weight–based doses of crovalimab once every 4 weeks via SC injection were considered as adequate and reasonable in clinical practice in Canada by the clinical experts. The clinical experts pointed out the importance of monitoring patients for any occurrence of kidney dysfunction or infections (particularly meningitis) during the treatment of crovalimab. According to the clinical experts and the clinician group, given that compliance may impact treatment effects, longer-term comparative evidence on the durability of the effectiveness of crovalimab would be informative. Likewise, the occurrence of some AEs, especially rare ones (e.g., meningitis), may take longer than 24 weeks to be identified. Comparative longer-term follow-up to assess safety between crovalimab and other complement inhibitors would be preferred.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For the pivotal studies and randomized controlled trials (RCTs) identified in the sponsor’s systematic review, the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework was used to assess the certainty of the evidence for outcomes considered most relevant to inform the expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.^12,13 Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, the imprecision of effects, and publication bias.

The selection of outcomes for GRADE assessment was based on the sponsor’s summary of clinical evidence, consultation with clinical experts, and input received from patient and clinician groups and public drug plans. The following list of outcomes was finalized in consultation with expert committee members:

• clinical outcomes — hemolysis control, BTH, mean percentage change in LDH levels, stabilized hemoglobin, having reached and maintained a minimum hemoglobin level, transfusion avoidance, and pRBC units transfused

• HRQoL outcomes — FACIT-F score and EORTC QLQ-C30 GHS and QoL score

• harms outcomes — infections and deaths.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty of evidence assessment was the presence or absence of an important effect based on the threshold (minimal important difference [MID]) in the literature for the FACIT-F score in both trials.

For the COMMODORE 2 study, the target of the certainty of evidence assessment was the presence or absence of a predefined NIM per study protocol for hemolysis control, BTH, stabilized hemoglobin, having reached and maintained a minimum hemoglobin level, and transfusion avoidance. Due to the lack of a formal MID estimate or NIM value, the target of the certainty of evidence assessment was the presence or absence of any (non-null) effect for the mean percentage change in LDH levels, pRBC units transfused, the EORTC QLQ-C30 GHS and QoL score, infections, and deaths.

For the COMMODORE 1 study, due to the lack of a formal MID estimate or NIM (after amendment of the COMMODORE 1 study, all efficacy end points became exploratory and descriptive, and thus NIM values were not applicable), the target of the certainty of evidence assessment was the presence or absence of any (non-null) effect for hemolysis control, BTH, mean percentage change in LDH levels, stabilized hemoglobin, having reached and maintained a minimum hemoglobin level, transfusion avoidance, pRBC units transfused, the EORTC QLQ-C30 GHS and QoL score, infections, and deaths.

For the GRADE assessments, the COMMODORE 2 and COMMODORE 1 studies were assessed individually because the COMMODORE 2 study included patients naive to complement inhibition and the COMMODORE 1 study included complement inhibition–experienced patients.

Results of GRADE Assessments

Table 2 and Table 3 present the GRADE summary of findings for crovalimab versus eculizumab for patients with PNH who were naive to and had experience with complement inhibitors, respectively.

Long-Term Extension Studies

Description of Studies

The phase I and phase II COMPOSER trial (ClinicalTrials.gov identification number NCT03157635) consisted of 4 sequential parts and an open-label extension (OLE). This section includes a summary of the OLE period that evaluated the safety, tolerability, PKs, and PDs of crovalimab in patients with PNH, aged 18 years to 75 years, who were treatment-naive or who switched from eculizumab. Participants were enrolled at 14 sites in 6 countries: Germany (3 sites), Japan (5 sites), France (1 site), Hungary (2 sites), Korea (2 sites), and Italy (1 site). The study did not include Canadian sites.

Efficacy Results

Hemolysis Control

The proportion of patients reaching an LDH level of 1.5 or less multiplied by ULN per visit suggested there was little to no difference during the OLE period, with 80% to 100% of the evaluable patients at each visit having an LDH level of 1.5 or less multiplied by ULN.

The mean normalized LDH level was generally maintained below 1.5 multiplied by ULN during the OLE phase. The point estimate of the mean normalized LDH level by visit ranged from 1.09 to 1.24 multiplied by ULN.

Table 2: Summary of Findings for Crovalimab Versus Eculizumab for Complement Inhibitor–Naive Patients With PNH

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; GHS = global health status; LDH = lactate dehydrogenase; MID = minimal important difference; NA = not applicable; NIM = noninferiority margin; NR = not reported; OR = odds ratio; PNH = paroxysmal nocturnal hemoglobinuria; pRBC = packed red blood cell; QoL = quality of life; RCT = randomized controlled trial; ULN = upper limit of normal.

Note: Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, the imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the Table 2 footnotes. Data presented in this table were based on analyses at the clinical cut-off date of November 16, 2022. The outcomes — patients who reached and maintained a minimum hemoglobin level, mean percentage change in LDH levels, mean units of pRBC transfused, and EORTC QLQ-C30 GHS and QoL score — were not adjusted for multiplicity and should be considered as supportive evidence.

^aImprecision did not result in the level of certainty being rated down. The point estimate and both the lower and upper boundaries of the 95% CI of the between-group comparison indicate trivial or no clinically meaningful difference according to the clinical experts consulted for this review. The 95% CI of the effect excludes the predefined NIM.

^bFor hemolysis control, the NIM for the lower 95% CI limit of the OR was 0.2. In the absence of available data for the between-group difference in the mean proportion of patients with hemolysis control, the judgment of imprecision is based on the 95% CI for the OR of crovalimab versus eculizumab using the predefined NIM of 0.2 as the threshold (the lower limit of the 95% CI was above 0.2, declaring noninferiority of eculizumab).

^cBreakthrough hemolysis was defined as at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, shortness of breath [dyspnea], anemia [hemoglobin < 10 g/dL], a major adverse vascular event [including thrombosis], dysphagia, or erectile dysfunction) in the presence of elevated LDH of 2 or more multiplied by ULN after a prior reduction of LDH to 1.5 or less multiplied by ULN on treatment. For breakthrough hemolysis, the NIM for the upper 95% CI limit for the weighted difference in proportions was 20%.

^dThe level of evidence was rated down 1 level for serious imprecision. There was no known MID and clinical experts consulted by CDA-AMC could not provide a threshold of important difference. The CDA-AMC review team judged that the upper bound of the 95% CI was close to the null and unlikely to include an important effect.

^eStabilized hemoglobin was defined as the avoidance of a 2 g/dL or more decrease in hemoglobin level from baseline, in the absence of transfusion. For stabilized hemoglobin, the NIM for the lower 95% CI limit of the weighted difference in proportions was −20%.

^fPatients who reached and maintained a minimum hemoglobin level were defined as patients who reached a hemoglobin level of at least 10 g/dL, without a subsequent decrease below 9 g/dL, in the absence of transfusion. For this outcome, the NIM was −20%.

^gTransfusion avoidance was defined as patients who were pRBC transfusion–free and who did not require a transfusion per protocol-specified guidelines. For transfusion avoidance, the NIM for the lower 95% CI limit of the weighted difference in proportions was −20%.

^hThe level of evidence was rated down 1 level for serious imprecision. Serious imprecision was defined as being unable to conclusively determine whether the between-group difference included the null threshold of 0, which would have been used to assess certainty given that the review team was unable to identify the MID to assess a between-group difference from the literature or the clinical experts consulted for this review. The within-group CIs overlapped and the overall sample size was small.

ⁱThis outcome presents the total number of pRBC units transfused among the full population.

^jThe level of evidence was rated down 1 level for serious risk of bias in favour of crovalimab (maintenance dose based on body weight every 4 weeks, SC injections) compared with eculizumab (maintenance dose of 900 mg every 2 weeks, IV infusion) arising from the open-label nature of the study and the subjective nature of the outcome.

^kFACIT-F was assessed in adult patients only (≥ 18 years). Imprecision did not result in the level of certainty being rated down. Based on the MID identified in the literature (5 points),¹⁴ which the clinical experts consulted for this review regarded as reasonable, the point estimate and both the lower and upper boundaries of the 95% CI of the between-group comparison suggested little to no difference.

^lThis analysis was not part of the sponsor's statistical analysis plan and was requested by CDA-AMC to facilitate a certainty of evidence appraisal.

^mThis outcome refers to the proportion of patients with at least 1 infection. No patients in either group had N. meningitidis. The level of evidence was rated down 2 levels for very serious imprecision. The review team was unable to identify the MID to assess a between-group difference from literature or the clinical experts consulted for this review; therefore, the null was used to assess certainty. The 95% CI of the absolute effect included the null threshold of 0.

ⁿThe level of evidence was rated down 2 levels for very serious imprecision. There was a very small number of events captured.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and sponsor’s submission.¹⁶

Table 3: Summary of Findings for Crovalimab Versus Eculizumab for Complement Inhibitor–Experienced Patients With PNH

CDA-AMC = Canada’s Drug Agency; CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; GHS = global health status; LDH = lactate dehydrogenase; MID = minimal important difference; NA = not applicable; NR = not reported; OR = odds ratio; PNH = paroxysmal nocturnal hemoglobinuria; pRBC = packed red blood cell; QoL = quality of life; RCT = randomized controlled trial; SC = subcutaneous; ULN = upper limit of normal.

Note: Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, the imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the Table 3 footnotes. Data presented in this table were based on analyses at the clinical cut-off date of November 16, 2022. The COMMODORE 1 trial was originally planned to recruit 200 patients to the randomized arms to be sufficiently powered to assess the efficacy of crovalimab. However, due to the introduction of ravulizumab to the treatment landscape and a reduced pool of patients treated with eculizumab over time, randomization to the crovalimab and eculizumab arms was stopped in November 2022 per protocol amendment version 6. The evaluation of safety became the primary objective and all efficacy objectives became exploratory. The results of the exploratory efficacy analyses were only reported descriptively because there was not sufficient power for formal statistical noninferiority or superiority testing. All outcomes in this table were not adjusted for multiplicity and should be considered as supportive evidence.

^aThe level of evidence was rated down 2 levels for very serious imprecision. The review team was unable to identify the MID to assess a between-group difference from literature or the clinical experts consulted for this review; therefore, the null was used to assess certainty. The 95% CI of the OR included the null threshold of 1.

^bBreakthrough hemolysis was defined as at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, shortness of breath [dyspnea], anemia [hemoglobin < 10 g/dL], a major adverse vascular event [including thrombosis], dysphagia, or erectile dysfunction) in the presence of an elevated LDH level of 2 or more multiplied by ULN after prior reduction of the LDH level to 1.5 or less multiplied by ULN on treatment.

^cThe level of evidence was rated down 2 levels for very serious imprecision. The review team was unable to identify the MID to assess a between-group difference from literature or the clinical experts consulted for this review; therefore, the null was used to assess certainty. The 95% CI of the absolute effect included the null threshold of 0.

^dStabilized hemoglobin was defined as the avoidance of a 2 g/dL or greater decrease in the hemoglobin level from baseline, in the absence of transfusion.

^ePatients who reached or maintained a minimum hemoglobin level were defined as patients who reached or maintained a hemoglobin level of at least 10 g/dL, without a subsequent decrease below 9 g/dL, in the absence of transfusion.

^fTransfusion avoidance was defined as patients who were pRBC transfusion–free and did not require a transfusion per protocol-specified guidelines.

^gThis outcome presents the total number of pRBC units transfused among the full population.

^hThe level of evidence was rated down 2 levels for very serious imprecision because the within-group CIs overlapped, the overall sample size was small, and the review team was unable to conclusively determine whether the between-group difference included the null threshold of 0, which would have been used to assess certainty given that the review team was unable to identify the MID to assess a between-group difference from the literature or the clinical experts consulted for this review.

ⁱThe level of evidence was rated down 1 level for serious risk of bias in favour of crovalimab (maintenance dose based on body weight every 4 weeks, SC injections) compared with eculizumab (900 mg every 2 weeks, IV infusion) arising from the open-label nature of the study and the subjective nature of the outcome.

^jFACIT-F was assessed in adult patients only (≥ 18 years). The level of evidence was rated down 1 level for serious imprecision. Based on the MID identified in the literature (5 points)¹⁴ that the clinical experts consulted for this review regarded as reasonable, the point estimate and the lower boundary of the 95% CI suggested little to no difference, and the 95% CI for the between-group difference crossed the MID threshold.

^kThe level of evidence was further rated down 1 level for a potential risk of bias due to missing data (14% in the crovalimab arm and 24% in the eculizumab arm). The impact of these missing outcome data is unclear.

^lThis analysis was not part of the sponsor's statistical analysis plan and was requested by CDA-AMC to facilitate a certainty of evidence appraisal.

^mThis outcome refers to the proportion of patients with at least 1 infection. No patients in either group had N. meningitidis. This analysis was not part of the sponsor's statistical analysis plan and was requested by CDA-AMC to facilitate a certainty of evidence appraisal. The level of evidence was rated down 2 levels for very serious imprecision. There was a very small number of events captured.

ⁿThe level of evidence was rated down 2 levels for very serious imprecision. There was a very small number of events captured.

Sources: COMMODORE 1 Primary Clinical Study Report¹⁷ and sponsor’s submission.¹⁶

Transfusion Avoidance

The proportion of patients with transfusion avoidance remained relatively stable over time during the OLE phase, with 82.9% to 91.7% of all patients avoiding transfusion across the 24-week intervals. Seven of the 11 patients who received a blood transfusion during the OLE phase reported a history of aplastic anemia and/or anemia.

Stabilized Hemoglobin

The proportion of patients reaching hemoglobin stabilization also remained relatively stable over time during the OLE phase, with 79.5% to 87.5% of all patients reaching hemoglobin stabilization across the 24-week intervals.

Breakthrough Hemolysis

The proportion of patients with BTH events was low during the OLE phase, with 0% to 4.9% of all patients reporting a BTH event across the 24-week intervals. One patient who had a BTH event discontinued the study before completing the full interval from week 20 to week 44, during which the BTH event occurred. Consequently, this BTH event was reported and included in the computation of the rate of BTH events adjusted for patient-years at risk but is not included in the calculation of the proportion of patients with BTH events per 24-week interval analyses.

Harms Results

███████████ of the 44 patients who were enrolled in part 2, part 3, and part 4 of the study entered the OLE phase, and █████ patients (██████ naive patients and ██████ patients who switched from eculizumab) were ongoing on crovalimab treatment at the time of the most recent clinical cut-off date (Table 23) (August 28, 2023). Overall, a total of ██████ patients discontinued from the OLE phase (██████ patient originally enrolled in part 2 [naive], █████ patients originally enrolled in part 3 [switch], ██████ patients originally enrolled in part 4, arm A [naive], and ██████ patient originally enrolled in part 4, arm B [switch]). No patients discontinued from the study due to an AE. The median treatment duration from baseline to the CCOD was ████ years.

A total of ██ ███████ patients experienced at least 1 AE from baseline to the CCOD. There was 1 death (in a patient who switched from eculizumab) that occurred in the OLE phase, in the period since the publication of the previous Update Clinical Study Report, and the death was not considered by the investigators to be related to study treatment. No AEs resulted in the withdrawal of a patient from the study. A total of ██ ███████ patients experienced at least 1 SAE. SAEs were reported in █████ of patients in the treatment-naive group and in █████ of patients in the switched-from-eculizumab group. The most frequently reported AEs by System Organ Class (with at least a 50% incidence in the total population) were infections and infestations (█████), gastrointestinal disorders (█████), and musculoskeletal and connective tissue disorders (█████).

The most frequently reported AEs with at least a 15% incidence in the total population were nasopharyngitis (█████), COVID-19 (█████), upper respiratory tract infection (█████), headache (█████), back pain (█████), pyrexia (█████), arthralgia (█████), cough (█████), bronchitis (█████), and dizziness (█████).

A total of ██ ███████ patients experienced at least 1 infection up to the CCOD (this was an increase from 36 [81.8%] patients as reported in the Update Clinical Study Report).

Critical Appraisal

The OLE period of the COMPOSER study was conducted to evaluate the longer-term safety and exploratory efficacy outcomes in 44 patients with PNH, beyond the 20-week primary evaluation period. Patients were enrolled in the extension period if study investigators determined that the patient derived benefit from treatment with crovalimab in the 20-week primary study period. The open-label design may have increased the risk of bias in patient selection and outcome ascertainment for end points that included more subjective assessments because the lack of blinding may have impacted investigators and patients’ expectations of treatment. The direction and magnitude of potential bias remains unclear. Given the design, the long-term extension was noncomparative because all patients in part 2, part 3, and part 4 received crovalimab. As a result, the comparative safety and efficacy of crovalimab to relevant comparators were not addressed and precluded conclusions about the comparative safety and efficacy of crovalimab. The small number of patients included in each study part, the single-arm setting, and the open-label design introduced significant uncertainty to study results and conclusions. There was variability in visit schedules across study parts, which may have introduced confounding to the study and limited the number of patients available at each follow-up time point.

The efficacy-evaluable population included all efficacy-evaluable patients who completed the primary treatment period in part 2, part 3, or part 4 of the study and were enrolled in the OLE period (18 treatment-naive patients and 25 switched patients). The study analysis included patients who were both treatment-naive and treatment-experienced — populations that differed in disease characteristics, such as transfusion history and LDH level. The impact of including both naive and experienced patients in the same analysis on the results is unclear.

During the OLE period, patients who were not receiving dosing every 4 weeks were required to switch to receive either 680 mg SC every 4 weeks (body weight ≥ 40 kg to < 100 kg) or 1,020 mg SC every 4 weeks (body weight ≥ 100 kg), which is aligned with the product monograph recommendation.

No method for the imputation of missing values was reported, and the attrition rate was 11.6%.

The study included diverse geographical representation and populations with varying demographics and health care systems, which may have enhanced the generalizability of the results. However, the included patients were mostly men (60% to 86%), with ethnicity of “not Hispanic or Latino,” and included patients with median ages ranging from 44 years to 56 years; therefore, results may not be generalizable to a broader population. Patients with conditions such as a psychiatric disorder, hereditary complement deficiency, a history of meningococcal meningitis, active primary or secondary immunodeficiency, a known history of HIV infection, chronic active hepatitis C, malignant disease, and alcohol, drug, or chemical abuse within 1 year before screening were excluded, and the results may not be generalizable to these groups of patients.

The dose-escalation approach, which allowed for the inclusion of different dosing frequencies, may have enhanced the generalizability across patient subgroups; however, frequent modifications and variations between groups increases intergroup variability and impacts the certainty and interpretation of study results.

Indirect Comparisons

Description of Studies

Due to the scarcity of direct evidence comparing crovalimab with other existing relevant therapies, including ravulizumab, eculizumab biosimilar, and the standard of care without a C5 inhibitor for PNH, the sponsor conducted a network meta-analysis (NMA) comparing crovalimab with ravulizumab, eculizumab biosimilar, and the standard of care without a C5 inhibitor for PNH.

Efficacy Results

The NMA suggested little to no difference between crovalimab and ravulizumab for the mean difference in the proportion of patients with transfusion avoidance, hemoglobin stabilization, and number of pRBC units transfused; however, the 95% credible intervals (CrIs) were wide and included the possibility that either treatment may be favoured. The NMA suggested little to no difference between crovalimab and ravulizumab for the change in the FACIT-F score; however, the CIs were wide and included the possibility of a clinically important effect that favoured crovalimab.

Harms Results

███ ███ ██████████ ██ ██████████ ████ ███████████ ███ ███ ████ ██ ████████████ ███ ███████ ███████ ███ █████ ████████ ████████ ████████████ ████████ ███ ███ ███ ███ ████ ███ ████████ ███ █████████ ████ ██████ █████████ ███ ██ █████████.

Critical Appraisal

Overall, the NMA was conducted according to accepted methodological guidance. A key limitation of the NMA was the heterogeneity in effect modifiers and prognostic factors across the included studies. The base-case analysis included patients who were both treatment-naive and treatment-experienced — populations that differed in disease characteristics such as transfusion history and LDH level. This heterogeneity suggests that the assumption of similarity across the included studies may not hold true for the NMA, increasing the uncertainty about the validity of the results for determining the effectiveness of crovalimab compared to ravulizumab. In a subgroup analysis of treatment-naive and treatment-experienced patients, the results were overall consistent with base-case results. Other limitations included the fact that the evidence networks were sparse, the connections between treatment nodes were informed by 1 or 2 trials, and the total number of patients in the included studies was relatively small. The aforementioned limitations may have increased the potential for biased treatment-effect estimates, and the results of the NMA should be interpreted with consideration of the limitations.

Studies Addressing Gaps in the Evidence From the Systematic Review

No studies addressing gaps in the pivotal and RCT evidence were identified for this review.

Conclusions

Two phase III, multicentre, open-label RCTs evaluated the efficacy and safety of SC injection of crovalimab at a weight-based dose once every 4 weeks compared with IV infusion of eculizumab 900 mg once every 2 weeks in adult patients with PNH after 24 weeks of treatment. Evidence from the COMMODORE 2 trial in patients with PNH who were naive to C5 inhibitors demonstrated that crovalimab was noninferior to eculizumab across the coprimary end points (hemolysis control [measured by LDH ≤ 1.5 × ULN] and transfusion avoidance) and secondary outcomes (BTH and stabilized hemoglobin). Other secondary (FACIT-F score) and exploratory end points were supportive of the noninferiority results, suggesting overall little to no difference between the treatment groups. Efficacy outcomes were rated as being of high to moderate certainty (except the EORTC QLQ-C30 GHS and QoL score, which was rated as low certainty), using the GRADE approach.

Results from the COMMODORE 1 trial in patients with PNH who were exposed to eculizumab were overall supportive of the results observed in the COMMODORE 2 trial, suggesting that crovalimab results in little to no difference in hemolysis control, BTH, transfusion avoidance, and the FACIT-F score compared with eculizumab. Efficacy analyses in the COMMODORE 1 trial were descriptive without formal statistical testing and should be considered as supportive evidence. The evidence was rated as being of low certainty, using the GRADE approach.

Descriptive analyses on long-term efficacy and safety based on the OLE phases of the COMMODORE 1 and COMMODORE 2 trials (an additional 24 weeks of treatment with crovalimab) and an OLE study (the COMPOSER study, with crovalimab treatment for a median duration of 3.40 years) appeared consistent with the randomized 24-week treatment periods of the pivotal trials for the crovalimab group, suggesting an ongoing benefit of crovalimab.

Based on the results of the COMMODORE 2 and COMMODORE 1 studies, the overall safety profile of crovalimab was consistent with that expected for a C5 inhibitor, and no additional safety signals were identified from the longer-term, single-arm COMPOSER study; however, long-term comparative data were not available. Approximately 1 in 5 patients who switched to receive crovalimab from eculizumab reported type III immune complex–mediated reactions.

The sample size of pediatric patients was small in the COMMODORE 2 study (n = 6) and in the COMMODORE 1 study (n = 1), and only descriptive results were available. The clinical experts anticipated that pediatric patients would have efficacy results similar to those of the main trial arms; however, there is a need for enhanced safety consideration regarding the risk of infections, including meningitis in pediatric patients.

Indirect evidence from the sponsor-submitted NMA suggested that crovalimab may provide efficacy and safety comparable to that of ravulizumab. Despite the analyses suggesting overall little to no difference between crovalimab compared to ravulizumab, the evidence was insufficient (i.e., a limited number of included studies, heterogeneity in patient characteristics across trials, and CrIs that crossed the null) to draw definitive conclusions on the relative efficacy of crovalimab compared to ravulizumab. The clinical experts consulted by the review team indicated that results suggesting little to no difference are plausible given that the 2 comparators belong to the same group of drugs and ravulizumab and eculizumab have demonstrated similar benefits in adult patients with PNH in RCTs.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of crovalimab (Piasky) 340 mg/2 mL for SC injection following the 1 loading dose by IV infusion in the treatment of PNH in adults and adolescents aged 13 years and older with a body weight of at least 40 kg.

Disease Background

Content in this section has been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the review team.

PNH is a rare, life-threatening, complement-mediated hematological disorder. PNH is the primary manifestation of chronic intravascular hemolysis, associated with bone marrow failure and thrombophilia.¹ It develops when hematopoietic cells acquire somatic mutations in the X-linked gene of PIGA.²

PNH can develop at any age, but it is most commonly diagnosed in the third decade of life, with no ethnic, geographic, or sex preferences.⁴ The mortality rate of patients with PNH who receive supportive care alone is approximately 20% to 35% within 5 years to 10 years of diagnosis.⁴ Chronic kidney disease, pulmonary hypertension, and venous or arterial thromboembolic events are life-threatening complications of PNH.⁴ Thromboembolic events were the leading cause of death in patients with PNH (40% to 67% of deaths with known cause) and their risk is high even among patients with no clinical evidence of thromboembolism or those under prophylactic anticoagulation therapy.⁵ In 1 study, the disease-specific cumulative incidence of mortality at 10 years was reported as 5.2%, with a transplant considered as a competing risk.⁵

Key clinical signs and symptoms of PNH include anemia, fatigue, thrombosis, esophageal spasms, male erectile dysfunction, hemoglobinuria, abdominal pain, and dyspnea.^1,4 Patients with PNH and their caregivers may experience impaired HRQoL and may struggle to complete normal everyday activities.¹⁸

Although epidemiological data are sparse, the previously estimated 15-year prevalence of PNH is 15.9 patients per 1 million population and the incidence is estimated at approximately 1.3 cases per million individuals per year.¹⁹ As of January 2018, the Canadian PNH Network followed 67 patients on therapy as well as 97 patients with small clones in the context of bone marrow failure; up to 52% of these patients were female and the median age of disease onset was 43 years.³

A diagnosis of PNH may be suspected if a patient presents with hemoglobinuria, cytopenia, thrombosis, aplastic anemia and/or myelodysplasia, or Coombs-negative hemolysis, with no known cause.³ A diagnosis of PNH is made by a combination of clinical examination, patient history, a physical exam, and specialized tests.³

A diagnosis of PNH will be confirmed by peripheral blood flow cytometry to detect the loss of glycosylphosphatidylinositol (GPI)-linked structures.³

The International PNH Interest Group has classified disease with GPI-deficient PNH clones into 3 subtypes: classic PNH, PNH in the context of other primary bone marrow disorders, and subclinical PNH.²⁰ In the classic PNH subcategory, there is clinical evidence of intravascular hemolysis; however, there is no evidence of another defined bone marrow abnormality. Patients with PNH in the context of another specified bone marrow disorder have clinical and laboratory evidence of hemolysis plus a history of marrow abnormality. In the last category, patients with subclinical PNH have no clinical or laboratory evidence of hemolysis, but there are small populations of hematopoietic cells deficient in GPI-anchored protein.²⁰

A Belgian expert panel recommended that patients with a granulocyte clone size of greater than 0.01% need to be monitored regularly, with frequency determined by clinical manifestations and clone size. In the absence of clinical manifestations, recommendations suggest that those with a clone size of less than 1% should be tested every year, whereas those with a clone size of greater than 1% should be tested every 6 months.²¹

Standards of Therapy

Content in this section has been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the review team.

The goals of PNH treatment are to reduce mortality, the number of thromboembolic events, and the risk of intravascular hemolysis; slow the progression of pulmonary pressures and renal impairment; alleviate PNH symptoms; improve patients’ HRQoL; and help patients regain better functional status, enabling a return to prediagnosis activities and employment.²² The clinical experts consulted for this review also pointed out that for pediatric patients with PNH, an additional treatment goal would be to control the enhanced complications risks of intravascular hemolysis and thromboembolic events because of aplastic or hypoplastic anemia, the condition that is often associated with PNH in children.

An allogeneic bone marrow transplant is considered a curative treatment for PNH.^6,7 However, a bone marrow transplant is not usually offered as a first-line treatment for classic PNH, owing to the risks of transplant-related morbidity and mortality, and is usually reserved for pediatric or younger adult patients with PNH with concomitant severe aplastic anemia or other causes of bone marrow failure.^4,8,9 The clinical experts consulted for this review noted that immunosuppressive therapy can be a treatment option when there is no bone marrow donor.

In Canada, patients with PNH receive C5 inhibitors as standard first-line therapy to reduce uncontrolled complement activation in the terminal complement cascade, and the primary C5 inhibitor therapies are ravulizumab (IV infusion administration every 8 weeks as a maintenance weight-based dose after a single loading weight-based dose) and eculizumab (IV infusion administration every 14 days as a maintenance dose after weekly doses over the first 5 weeks after initiation).⁴ The clinical experts consulted for this review noted the treatment algorithm for patients with PNH in Canada as follows: once a patient has developed the criteria for high disease activity (an elevated LDH level > 1.5 × ULN and 1 of the symptoms of fatigue, thrombosis, renal insufficiency, transfusions, and anemia), an application for eculizumab or ravulizumab would be submitted to the provincial or territorial special authorization drug plan. The current funding criteria for eculizumab vary across provinces and territories, but are generally aligned to initiate complement inhibition in patients with a leukocyte PNH clone of greater than 10%, laboratory evidence of significant intravascular hemolysis, and at least 1 of the following: symptomatic anemia (regardless of transfusion dependence), thrombosis, renal insufficiency, pulmonary insufficiency or hypertension, abdominal pain requiring admission, or opioid analgesia.³ The clinical experts consulted for this review noted that the majority of patients would start on C5 inhibition in the upfront setting, with ravulizumab being preferred due to the convenience to the patient and a lower risk of BTH, but eculizumab is likely preferred in patients who are pregnant. Eculizumab and immunosuppressive therapy can be used concurrently in situations where both are felt to be indicated.²³ Of note, ravulizumab is not reimbursed in all Canadian jurisdictions (e.g., British Columbia), leaving eculizumab as the only C5 inhibitor option for some patients.²³

In Canada, patients receiving C5 inhibitors who are diagnosed with extravascular hemolysis may be switched to pegcetacoplan.^24,25 Pegcetacoplan, a proximal C3 inhibitor, was approved by Health Canada in December 2022 and is indicated for the treatment of adults with PNH who have an inadequate response to, or are intolerant of, a C5 inhibitor.²⁶ It is administered subcutaneously twice a week via a syringe system infusion pump.²⁵ Pegcetacoplan has been previously reviewed by CDA-AMC with a recommendation of reimburse with conditions (April 2023).²⁴

Supportive care may be necessary for some patients with PNH, comprising additional transfusions, iron and folic acid supplementation, and analgesia to manage abdominal pain and smooth muscle dystonia.⁴

Drug Under Review

Key characteristics of crovalimab are summarized in Table 4 with other treatments available for the treatment of patients with PNH.

Crovalimab solution, 340 mg/2 mL (170 mg/mL) for injection and infusion, is indicated for the treatment of PNH in adults and adolescents aged 13 years and older with a body weight of at least 40 kg. It has not been previously reviewed by CDA-AMC.

The recommended dosing regimen consists of 1 loading dose administered by IV infusion (on day 1), followed by 4 additional weekly loading doses administered by SC injection (on day 2, day 8, day 15, and day 22). The maintenance dose starts on day 29 and is then administered every 4 weeks by SC injection.

Loading dose:

• On day 1, the dose is 1,000 mg (IV) for a body weight of 40 kg to less than 100 kg, and 1,500 mg (IV) for a body weight of 100 kg or more.

• On day 2, day 8, day 15, and day 22, the dose is 340 mg (SC) for a body weight of 40 kg or more.

Maintenance dose: On day 29 and every 4 weeks thereafter, the dose is 680 mg (SC) for a body weight of 40 kg to less than 100 kg, and 1,020 mg (SC) for a body weight of 100 kg or more.

Crovalimab is a recombinant humanized immunoglobulin G1–based monoclonal antibody that specifically binds with high affinity to component 5 of the complement system (C5), inhibiting its cleavage into C5a and C5b and thus preventing the formation of the membrane attack complex. In patients with PNH, crovalimab inhibits terminal complement-mediated intravascular hemolysis. Crovalimab binds to a different C5 epitope compared to other C5 inhibitors ravulizumab and eculizumab.

The sponsor’s reimbursement request is for the treatment of PNH in adults and adolescents aged 13 years and older with a body weight of at least 40 kg; this is aligned with the Health Canada indication.

Crovalimab is approved by FDA for the treatment of adult and pediatric patients 13 years and older with PNH and a body weight of at least 40 kg. In the European Union, crovalimab as monotherapy is indicated for the treatment of adult and pediatric patients aged 12 years or older with a weight of 40 kg or more with PNH:

• in patients with hemolysis with clinical symptom(s) indicative of high disease activity

• in patients who are clinically stable after having been treated with a C5 inhibitor for at least the past 6 months.

In the European Union, crovalimab is under additional monitoring, meaning that it is monitored even more intensively than other medicines.²⁷

Table 4: Key Characteristics of Crovalimab, Eculizumab, and Ravulizumab

PNH = paroxysmal nocturnal hemoglobinuria; SC = subcutaneous.

^aHealth Canada–approved indication.

Sources: Product monograph for crovalimab (Piasky),²⁸ product monograph for eculizumab (Soliris),¹¹ and product monograph for ravulizumab (Ultomiris).²⁹

Perspectives of Patients, Clinicians, and Drug Programs

The full patient and clinician group submissions received by CDA-AMC are available in the consolidated patient and clinician group input document for this review on the project website.

Patient Group Input

This section was prepared by the review team based on the input provided by patient groups.

CDA-AMC received 1 joint input from the Canadian Association of PNH Patients and AAMAC. The Canadian Association of PNH Patients is a nonprofit Canadian organization dedicated to advocating for optimal patient care, ensuring access to the latest tools and information for managing the condition effectively, raising awareness, and offering support to caregivers. AAMAC’s mission is to offer comprehensive support to people in Canada affected by aplastic anemia, myelodysplasia, or PNH, including patients, family members, friends, and health care providers.

Due to the limited size of the clinical trial for crovalimab, the patient groups were able to receive input from only 1 patient living in Canada. This patient had experience with crovalimab treatment. In this submission, the patient groups also provided information from the crovalimab clinical trial to highlight the trial’s findings and demonstrate the potential efficacy of crovalimab.

The patient groups stated that PNH is a complex and multifaceted disorder that requires comprehensive management to address the various aspects of the disease and improve the QoL of those affected. The chronic nature of PNH means that patients must manage their condition over a lifetime, dealing with the physical, emotional, and financial burdens associated with the disease. The impact on QoL is profound, as patients must cope with the unpredictability of symptoms, the side effects of treatments, and the constant threat of serious complications. Regular monitoring and supportive care are essential to managing the disease and improving patient outcomes.

According to the patient group input, 1 patient with PNH in Canada reported that the disease started with feeling constantly exhausted, followed by dark urine in the morning and chest pain. The patient explained feeling terrified of the risk of blood clots, and not knowing the effect of the disease on life, job, and family, along with having to accept that this is a lifelong condition.

Based on the input, 1 significant challenge with current treatments is that they require regular clinic visits every 2 weeks to 8 weeks for IV infusions, which can be time-consuming and disruptive. Because crovalimab can be given as a small SC injection, it is more convenient for patients than IV treatments. Based on the patient group input, the patient with PNH reported that the SC crovalimab treatment that only requires administration once every 4 weeks was much more convenient than a treatment that requires regular IV infusions. Furthermore, the patient reported that the SC injections are quick and easy, and can be administered at home. This patient reported that crovalimab significantly reduced hemolysis, improved energy levels within the first few months, reduced fatigue, increased independence, and alleviated the fear of blood clots. The patient also noted that the side effects of crovalimab were minimal.

The patient groups stated that crovalimab has the potential to greatly improve QoL. According to the patient groups, crovalimab reduces the physical, emotional, and logistical burdens associated with traditional IV therapies, allowing patients to regain a sense of freedom and control over their lives. Fewer clinic visits and less travel for IV injections mean lower health care costs and less financial burden for patients. The more convenient and less invasive approach can significantly improve not only patients’ QoL but also patients’ overall well-being and adherence to treatment, leading to better health outcomes.

Clinician Input

Input From Clinical Experts Consulted for This Review

All CDA-AMC 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, providing guidance on the potential place in therapy). The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of PNH.

Unmet Needs

The clinical experts consulted for this review noted that crovalimab may fulfill an unmet need for some patients with a preference for an SC route of drug administration, particularly for those with poor venous access and who are opposed to the insertion of a port-a-cath (a tube that is inserted into a vein and connected to a port that is placed under the skin). The clinical experts noted that crovalimab may be a safer alternative for patients living in isolated geographic locations that preclude the guaranteed delivery of eculizumab, and some patients may be more comfortable with self-administration than having to travel to a clinic for IV therapy (which is associated with transportation costs and exposure to potential infections at a clinic).

Place in Therapy

According to the clinical experts consulted for this review, most patients with PNH would choose to start ravulizumab as the first-line therapy over eculizumab or crovalimab where ravulizumab is funded (of note, ravulizumab is not funded in some jurisdictions such as British Columbia). However, eculizumab would be preferred for patients who are pregnant. Furthermore, the clinical experts noted that patients who prefer to be more independent in receiving therapy, have significant needle phobia, or find it difficult to visit a clinic due to geographic locations may choose crovalimab as first-line therapy. The clinical experts noted that switching from ravulizumab or eculizumab to crovalimab could be uncommon because there is a risk of transient immune complex formation that only occurs in patients switching to or from other C5 inhibitors.

The clinical experts further pointed out that in children with PNH, due to the high association with aplastic or hypoplastic anemia, the management would be driven by treating aplastic or hypoplastic anemia and allogeneic hematopoietic stem cell transplant is commonly considered a treatment option in children with PNH. The clinical experts noted that the rare cases of isolated PNH would be treated along the same lines as those of adults, except that PNH clone size is smaller in children and only clones of greater than 10% would be eligible for C5 inhibition therapy. The clinical experts noted that crovalimab would be indicated in the rare setting of PNH without aplastic anemia.

Patient Population

The clinical experts noted that patients who are fully motivated to self-administer the SC injections and those known to have genetic polymorphism that obviates efficacy with eculizumab or ravulizumab would be best suited for treatment with crovalimab.

Assessing the Response to Treatment

The clinical experts pointed out that crovalimab is another C5 inhibitor. Thus, the outcomes used to assess the treatment response would be the same as for eculizumab or ravulizumab, including improvements in hemoglobin levels, transfusion avoidance, and renal function, along with reductions in smooth muscle spasm, less fatigue, the normalization of LDH levels, and avoidance of BTH. The clinical experts noted that frequent laboratory and patient assessments are required during the early stages of crovalimab initiation, followed by standard assessments once the laboratory results stabilize.

Discontinuing Treatment

The clinical experts noted that if patients treated with crovalimab are not showing any clinical responses by 3 weeks to 4 weeks, a transition back to ravulizumab or eculizumab monitoring for the immune responses should be considered. The clinical experts pointed out that crovalimab should be discounted when any thrombosis occurs, or when there is an ongoing LDH elevation or ongoing symptoms of PNH.

Prescribing Considerations

The clinical experts commented that ideally, crovalimab should be prescribed by hematologists treating patients with PNH and the management of patients receiving crovalimab should take place at a centre of excellence.

Clinician Group Input

This section was prepared by the review team based on the input provided by clinician groups.

One clinician group input was received from the Canadian PNH Network. The Canadian PNH Network is a group of hematologists in Canada with a special interest and expertise in the care of patients with PNH; the members represent centres of expertise from Newfoundland and Labrador, Nova Scotia, Quebec, Ontario, Alberta, and British Columbia. Information for this submission was obtained via publicly available documents, congress abstracts, and the published literature, and members of the Canadian PNH Network were invited to contribute to the various sections. A total of 3 clinicians contributed to this submission.

According to the clinician group input, the current standard of care for patients with hemolytic PNH is terminal complement inhibition with a C5 blockade; as such, eculizumab and ravulizumab remain the only first-line therapies across Canada. The only curative treatment for PNH at this time is an allogeneic hematopoietic stem cell transplant. It should be noted, however, that this is reserved for patients with predominant or progressive bone marrow failure (e.g., aplastic anemia), which can coincide with, precede, or follow a diagnosis of PNH. More recently, danicopan received a positive conditional CDA-AMC recommendation (November 2024); it is used in combination with a continued C5 inhibitor, either eculizumab or ravulizumab. The CDA-AMC review team notes that at the time of receiving the crovalimab submission, danicopan was still under review with CDA-AMC.

The Canadian PNH Network highlighted that crovalimab is expected to address important treatment goals unmet by current therapies, including offering a C5 inhibitory strategy that does not require IV access, enabling self-administration by patients or administration by caregivers, and providing an option for rare cases of resistance to eculizumab or ravulizumab due to a C5 polymorphism. Additionally, the clinician group expected that crovalimab would be another first-line option for patients with PNH.

In terms of patients who are best suited for treatment with crovalimab, the Canadian PNH Network noted that crovalimab would be offered as first-line treatment to patients with PNH, either switching from an IV C5 inhibitor or starting from being treatment- naive. Patients who favour the freedom and reduced treatment burden of SC administration would very likely select crovalimab therapy. On the other hand, patients with PNH least suitable would be those who are not accepting SC drug delivery, or those who develop clinically significant extravascular hemolysis, which necessitates proximal complement inhibition strategies.

The Canadian PNH Network stated that response to treatment focuses on LDH reduction, which reduces hemolysis and the risk of thrombosis and may improve hemoglobin as well as transfusion dependence in patients with PNH. Clinical outcomes related to the response included decreased fatigue and transfusion requirements as well as improved QoL and overall survival. Efficacy outcomes would typically be assessed every 2 weeks to 4 weeks initially after starting a new therapy or switching, but follow-up would be required less often (e.g., every 3 months to 6 months) as a patient becomes established on the drug and does not show evidence of side effects or other concerns.

Based on the input, some of the factors that should be considered to discontinue the treatment include AEs, type III immune complex reactions, poor compliance, and pregnancy. The clinician group added that treatment with crovalimab is most likely going to be done either entirely at the patient’s home or, in the case of loading doses, at a local infusion clinic.

Drug Program Input

The drug programs provide input on each drug being reviewed through the reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CDA-AMC for this review are summarized in Table 5.

Table 5: Summary of Drug Plan Input and Clinical Expert Response

BIA = budget impact analysis; CDEC = Canadian Drug Expert Committee; DTDC = drug-target-drug complex; NIHB = Non-Insured Health Benefits; PNH = paroxysmal nocturnal hemoglobinuria; PT = province and territory; VAC = Veterans Affairs Canada.

Note: In the Canadian Expert Drug Advisory Committee final recommendation for eculizumab (Soliris),³¹ the recommendation was do not list at the submitted price. In CADTH Reimbursement Recommendation: Ravulizumab (Ultomiris),³² the Canadian Drug Expert Committee recommends that ravulizumab be reimbursed for the treatment of adult patients with PNH with clinical criteria and/or conditions, including for the treatment’s initiation, renewal, discontinuation, and prescribing. The clinical criteria and/or conditions are as follows: 1) list in a similar manner to eculizumab for initiation, renewal, discontinuation, and prescribing, with the addition of condition 2 for initiation and condition 3 for prescribing; 2) patients with insufficient initial response or who did not experience improvement with treatment with eculizumab at the Health Canada–recommended dosage are not eligible for reimbursement of ravulizumab; 3) ravulizumab should only be prescribed at the Health Canada–recommended dosage; 4) ravulizumab should be negotiated so that it does not exceed the drug program cost of treatment with eculizumab reimbursed for the treatment of adult patients with PNH; and 5) the feasibility of the adoption of ravulizumab must be addressed.

Clinical Evidence

The objective of this Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of crovalimab 340 mg/2 mL for SC injection following 1 loading dose by IV infusion in the treatment of PNH in adults and adolescents aged 13 years and older with a body weight of at least 40 kg. The focus will be placed on comparing crovalimab to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of crovalimab is presented in 3 sections with CDA-AMC critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. Assessment by CDA-AMC of the certainty of the evidence in this first section using the GRADE approach follows the critical appraisal of the evidence. The second section includes sponsor-submitted long-term extension studies. The third section includes indirect evidence from the sponsor. No additional studies were submitted to address important gaps in the systematic review evidence.

Included Studies

Clinical evidence from the following is included in the review and appraised in this document:

• 2 pivotal studies (RCTs) identified in systematic review

• 1 long-term extension study

• 1 indirect treatment comparison (ITC).

Systematic Review

Content in this section has been informed by materials submitted by the sponsor. The following has been summarized and validated by the review team.

Description of Studies

Characteristics of the included studies are summarized in Table 6.

The COMMODORE 2 and COMMODORE 1 studies are pivotal phase III RCTs evaluating the safety and efficacy of crovalimab compared to eculizumab in patients with PNH who had not been previously treated with a complement inhibitor (the COMMODORE 2 study) and patients with documented treatment with complement inhibitors (the COMMODORE 1 study).

COMMODORE 2 Study

The COMMODORE 2 trial was a multicentre, randomized, open-label, active-controlled study comparing the efficacy and safety of crovalimab compared with eculizumab in patients diagnosed with PNH who had not been previously treated with a complement inhibitor therapy, and who had a body weight of 40 kg or more. Adult patients were randomized 2:1 to receive crovalimab (arm A, n = 135) or eculizumab (arm B, n = 69). A central randomization procedure was used, where a block-based randomization was stratified by the most recent locally measured LDH level (2 to 4 × ULN, and > 4 × ULN) before randomization and number of pRBC units (0, > 0 to ≤ 6, and > 6) administered within the 6 months before randomization. Six additional patients were enrolled in arm C (descriptive arm that received crovalimab), which enrolled pediatric patients with PNH who were younger than 18 years. The primary treatment period was 24 weeks for all arms. After 24 weeks, all patients had the opportunity to continue or switch to crovalimab in the extension period (Figure 1). Patients were enrolled across 67 sites in 25 countries. The study did not include sites in Canada. The CCOD as submitted in the aligned review to Health Canada was November 16, 2022.^15,35

Table 6: Details of Studies Included in the Systematic Review

AE = adverse event; ANC = absolute neutrophil count; CTCAE v5 = Common Terminology Criteria for Adverse Events, Version 5; DTDC = drug-target-drug complex; EORTC = European Organisation for Research and Treatment of Cancer; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; Hb = hemoglobin; ICF = informed consent form; IPSS-R = Revised International Prognostic Scoring System; LDH = lactate dehydrogenase; MAVE = major adverse vascular event; MFS = multidimensional fatigue scale; NA = not applicable; NCI = National Cancer Institute; PedsQL = Pediatric Quality of Life Inventory; PGI-S = Patient Global Impressions–Severity; PNH = paroxysmal nocturnal hemoglobinuria; pRBC = packed red blood cell; q.2.w. = every 2 weeks; q.4.w. = every 4 weeks; QLQ-AA/PNH = Quality of Life Questionnaire – Aplastic Anemia and/or Paroxysmal Nocturnal Hemoglobinuria; RCT = randomized controlled trial; SC = subcutaneous; TSQM-9 = Treatment Satisfaction Questionnaire for Medication 9; ULN = upper limit of normal; WBC = white blood cell; vs. = versus.

Note: LDH level was assessed by the central laboratory in the COMMODORE 2 and COMMODORE 1 trials. Transfusion avoidance was defined as patients who are pRBC transfusion–free and who do not require a transfusion per protocol-specified guidelines. Breakthrough hemolysis was defined as at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, dyspnea, anemia [Hb < 10 g/dL], a MAVE, dysphagia, or erectile dysfunction) in the presence of an elevated LDH level of 2 or more multiplied by ULN after the prior reduction of the LDH level to 1.5 or less multiplied by ULN on treatment. Stabilized hemoglobin was defined as the avoidance of a 2 g/dL or greater decrease in hemoglobin level from baseline, in the absence of transfusion. Two additional reports were included (Roth et al. [2024]³³ and Scheinberg et al. [2024]³⁴).

^aTwo pediatric patients were randomized to eculizumab in arm B, aged 17 years.

^bThis exclusion criterion was added as an exclusion criterion to the COMMODORE 2 trial following the protocol amendment on January 24, 2022.

^cThe primary objective of the study was evaluating the safety and tolerability of crovalimab compared to eculizumab.

^dClinical outcomes were listed. Pharmacokinetic and pharmacodynamic outcomes were not listed in Table 6.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and COMMODORE 1 Primary Clinical Study Report.¹⁷

Figure 1: COMMODORE 2 Study Design

LDH = lactate dehydrogenase; PNH = paroxysmal nocturnal hemoglobinuria; RBC = red blood cell; ULN = upper limit of normal.

^aRandomization was stratified based on the most recent LDH value (2 to 4 × ULN, and > 4 × ULN) and packed RBC transfusion history (0 unit, > 0 to 6 units, and > 6 units) within 6 months before randomization. Patients were randomized 2:1 to crovalimab or eculizumab.

^bPatients who discontinued crovalimab entered the safety follow-up period of 46 weeks. Patients who discontinued eculizumab entered the safety follow-up period of 10 weeks.

^cAfter completing 24 weeks of treatment, patients randomized to eculizumab could continue treatment or switch to crovalimab. If they discontinued treatment, the safety follow-up period for eculizumab patients was 10 weeks.

Source: Sponsor’s submission.¹⁶

COMMODORE 1 Study

The COMMODORE 1 study is a multicentre, randomized, open-label, active-controlled study to evaluate the safety, pharmacokinetics (PKs), pharmacodynamics (PDs), and efficacy of crovalimab compared with eculizumab in patients diagnosed with PNH who were treated with a complement inhibitor therapy, and who had a body weight of 40 kg or more. Patients were enrolled across 70 sites in 25 countries, including 1 site in Canada. The COMMODORE 1 study was originally planned to recruit 200 patients to the randomized arms to be sufficiently powered to assess the efficacy of crovalimab. However, due to the introduction of ravulizumab to the treatment landscape and a reduced pool of patients treated with eculizumab over time, randomization to arm A and arm B was stopped in November 2022 per protocol amendment version 6.¹⁰ The evaluation of safety became the primary objective and all efficacy objectives became exploratory. As a result, 89 adult participants were randomized 1:1 to receive crovalimab (arm A, n = 45) or eculizumab (arm B, n = 44) in the modified design of the COMMODORE 1 study. A central randomization procedure was used. Randomization was stratified according to a patient’s transfusion history (received a transfusion of pRBCs within 12 months before randomization [yes or no]) using a block-based method. After completing 24 weeks of crovalimab or eculizumab treatment, all patients had the opportunity to continue or switch to crovalimab in the extension period (Figure 2).^10,17

The nonrandomized arm C originally intended to recruit about 50 patients in different clinically relevant cohorts. Following the protocol amendment to stop randomization due to a reduced pool of patients treated with eculizumab, adult patients (≥ 18 years) who had been receiving eculizumab at the approved dose for at least 24 weeks before study entry (i.e., the same population as those who would have been enrolled in the randomized arms) were able to enter arm C to continue access to crovalimab in a nonrandomized setting. At the CCOD of primary analysis, 38 patients were enrolled in arm C: 1 patient in the pediatric cohort, 21 patients in the prior ravulizumab cohort, 10 patients in the prior eculizumab at a higher-than-approved dose cohort, and 6 patients in the C5 polymorphism cohort.^10,17 Enrolment in arm C is ongoing.

The CCOD of the COMMODORE 1 study as submitted in the aligned review to Health Canada was November 16, 2022, which was the same date as for the COMMODORE 2 study.^10,17

Populations

Inclusion and Exclusion Criteria

COMMODORE 2 Study

Eligible patients had clinically significant disease activity at screening, demonstrated by an LDH level of 2 or more multiplied by ULN and the presence of 1 or more PNH-related signs or symptoms in the past 3 months. Patients with current or previous treatment with a complement inhibitor were excluded. Vaccination against N. meningitidis serotypes A, C, W, and Y less than 3 years before the initiation of study treatment was required. Randomized arm A and arm B enrolled adult patients, and descriptive arm C enrolled patients younger than 18 years at the time of signing the informed consent form.

Figure 2: COMMODORE 1 Study Design

PNH = paroxysmal nocturnal hemoglobinuria; pRBC = packed red blood cell; q.2.w. = every 2 weeks; SNP = single nucleotide polymorphism.

^aStratification: History of pRBC transfusion within 1 year before randomization (yes or no).

^bPatients who discontinued eculizumab or crovalimab entered a safety follow-up period of 10 weeks for eculizumab and 46 weeks for crovalimab.

^cAfter completing 24 weeks of treatment, patients randomized to eculizumab could continue treatment or switch to crovalimab.

^dFollowing closure of the randomized arms to recruitment, the nonrandomized arm was opened to adults receiving eculizumab at the approved dose for 24 weeks or more before study entry.

^eHigher-than-approved doses of eculizumab: More than 900 mg per dose and/or more frequently than q.2.w.

Source: Sponsor’s submission.¹⁶

COMMODORE 1 Study

Eligibility criteria in the COMMODORE 1 study (randomized arm A and arm B) are identical to those of the COMMODORE 2 study with the following exceptions:

• patients must have documented treatment with eculizumab according to the approved dosing recommended for PNH (900 mg every 2 weeks) and completion of a minimum of 24 weeks of treatment before day 1

• patients must have an LDH level of 1.5 or less multiplied by ULN at screening (indicating hemolysis control on current therapy, enabling a stable switch in the randomized period)

• patients are excluded if they have a pre-enrolment hemoglobin value of 7 g/dL or less, or a pre-enrolment hemoglobin value of more than 7 g/dL and up to 9 g/dL with concurrent signs and symptoms of anemia (of note, this exclusion criterion was also added as an exclusion criterion to the COMMODORE 2 trial following the protocol amendment on January 24, 2022).

Eligibility criteria in arm C of the COMMODORE 1 trial differs from those of arm C of the COMMODORE 2 trial, as outlined in Table 6.

Interventions

COMMODORE 2 and COMMODORE 1 Studies

Crovalimab was administered with the same dosing regimen in both the COMMODORE 2 and COMMODORE 1 studies. This is also the same dosing regimen as in the reimbursement request and the submission to Health Canada (Table 1 and Table 6).^23,28

In both studies, the crovalimab group received a weight-based tiered dosing regimen of crovalimab consisting of a loading series (an IV dose on day 1 [1,000 mg for a body weight of 40 kg to < 100 kg or 1,500 mg for a weight of ≥ 100 kg] followed by SC injection doses on day 2, day 8, day 15, and day 22 [340 mg]) and maintenance dosing (SC injection every 4 weeks starting on day 29 [680 mg for a body weight of 40 kg to < 100 kg or 1,020 mg for a weight of ≥ 100 kg]). For patients 40 kg to less than 100 kg receiving an initial IV loading dose of 1,000 mg, the infusion was delivered over 60 (± 10) minutes. For patients receiving an initial IV loading dose of 1,500 mg, the infusion was delivered over 90 (± 10) minutes. Patients were observed by a health care professional during the IV infusion and for 60 minutes following the completion of the IV infusion. Crovalimab self-administration (in patients ≥ 12 years only) or administration by a caregiver was permitted starting at week 9, after training and confirmation of proficiency by the health care professional. Patients who did not wish to self-inject or did not have assistance could continue to have crovalimab administered by the study site staff. If a patient experienced signs and symptoms of their underlying PNH, such as BTH (which may be due to an acute event such as acute illness, trauma or surgery), 1 or more additional IV doses of crovalimab could be administered as a rescue dose (340 mg IV regardless of body weight infused over 30 minutes).^15,17

The eculizumab group received eculizumab per local guidelines or per pharmacy manual: induction doses of 600 mg on day 1, day 8, day 15, and day 22, followed by maintenance doses of 900 mg on day 29 and every 2 weeks thereafter. No eculizumab dose modifications were permitted in the COMMODORE 2 and COMMODORE 1 studies.^15,17 Patients who discontinued treatment entered a safety follow-up period.

Concomitant Medications

In both the COMMODORE 2 and COMMODORE 1 studies, patients were permitted to use the following therapies: oral contraceptives with a failure rate of less than 1% per year, immunosuppressant therapy, corticosteroids, iron supplements, and folic acid.^10,35

Patients could receive other concomitant medication, which was recorded. Investigators could provide supportive therapies as clinically indicated, per local standard practice. Infusion-associated symptoms could be treated symptomatically with acetaminophen, ibuprofen, diphenhydramine, and/or histamine H₂ receptor antagonists (e.g., famotidine, cimetidine), or equivalent medications per local standard practice. Serious infusion-associated events manifested by dyspnea, hypotension, wheezing, bronchospasm, tachycardia, reduced oxygen saturation, or respiratory distress were managed with supportive therapies as clinically indicated (e.g., supplemental oxygen and beta₂-adrenergic agonists).^15,17

Outcomes

A list of efficacy end points assessed in this Clinical Review Report is provided in Table 7, followed by descriptions of the outcome measures. Summarized end points are based on outcomes included in the sponsor’s summary of clinical evidence as well as any outcomes identified as important to this review according to the clinical experts consulted for this review and input from patient and clinician groups and public drug plans. Using the same considerations, CDA-AMC selected end points that were considered to be most relevant to informing expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using the GRADE tool. Select notable harms outcomes considered important for informing expert committee deliberations were also assessed using GRADE.

Table 7: Outcomes Summarized From the Studies Included in the Systematic Review

AE = adverse event; AESI = adverse event of special interest; EORTC IL-40 = European Organisation for Research and Treatment of Cancer – Item Library 40; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; GHS = global health status; LDH = lactate dehydrogenase; NR = not reported; PPQ = Patient Preference Questionnaire; pRBC = packed red blood cell; QoL = quality of life; SAE = serious adverse event; ULN = upper limit of normal.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and COMMODORE 1 Primary Clinical Study Report.¹⁷

Efficacy Outcomes

The COMMODORE 2 and COMMODORE 1 trials investigated the same outcomes at week 25 (i.e., after a 24-week treatment period). A description of the efficacy outcome measures and the measurement properties that were used in both the COMMODORE 2 and COMMODORE 1 trials are presented in this section (Table 8).

Hemolysis Control

Serum LDH is a biochemical marker of intravascular hemolysis. The cut-off for a clinically meaningful threshold of improvement in both the COMMODORE 2 and COMMODORE 1 studies was an LDH level of 1.5 or less multiplied by ULN (as measured at the central laboratory), based on the available literature supporting its clinical relevance as the most sensitive and specific predictor of mortality and thromboembolic events.^36-39

Breakthrough Hemolysis

BTH was defined as at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, shortness of breath [dyspnea], anemia [hemoglobin < 10 g/dL], a major adverse vascular event (MAVE) [including thrombosis, dysphagia, or erectile dysfunction]) in the presence of an elevated LDH level of 2 or more multiplied by ULN after a prior reduction of the LDH level to 1.5 or less multiplied by ULN on treatment. PRO instruments were self-administered at the clinic at specified time points during the studies.

Stabilized Hemoglobin

Stabilized hemoglobin was defined as avoidance of a 2 g/dL or greater decrease in hemoglobin level from baseline, in the absence of transfusion. Baseline hemoglobin was defined as the latest available hemoglobin measurement before the first on-study drug administration of the study drug.

Transfusion Avoidance

Transfusion avoidance is a disease-related event for patients with PNH, defined as patients who are pRBC transfusion–free and do not require transfusions per protocol-specified guidelines. In both trials, transfusions were administered per the investigators' clinical judgment. In general, pRBC transfusions were recommended when a patient had hemoglobin of 9 g/dL or less with clinical signs and symptoms of sufficient severity to warrant transfusion or hemoglobin of 7g/dL or less regardless of clinical signs or symptoms.

FACIT-F (Version 4)

FACIT-F is a validated, reliable, self-report measure for use in a variety of conditions, including anemia.^40,41 The content validation research supports its use in patients with PNH, as it comprehensively covers PNH-related fatigue.⁴² The FACIT-F consists of 13 items that assess fatigue using a 7-day recall period. Items are scored on a response scale that ranges from 0 (“not at all”) to 4 (“very much so”). Relevant items are reverse-scored and all items are summed to create a total score ranging from 0 to 52, with higher scores indicating lower fatigue.^43,44 The threshold for a clinically meaningful change is 5 points or more.⁴⁴ FACIT-F was administered to adult patients only in both the COMMODORE 2 and COMMODORE 1 studies.

EORTC QLQ-C30 (Version 3)

EORTC QLQ-C30 is a validated, reliable, self-report measure.⁴⁵ Although the measure was originally developed for use in patients undergoing cancer treatment, recent content validation research supports its use in patients with PNH.⁴² It consists of 30 questions that assess 5 aspects of patient functioning (physical, emotional, role, cognitive, and social), 3 symptom scales (fatigue, nausea and vomiting, and pain), GHS and QoL, and 6 single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties), with a recall period of 1 week. Scale scores can be obtained for the multi-item scales. The functioning and symptom items are scored on a 4-point scale that ranges from “not at all” to “very much,” and the GHS and QoL items are scored on a 7-point scale that ranges from “very poor” to “excellent.” This PRO tool was administered in a truncated version of the measure that included only the physical functioning, role functioning, and GHS and QoL scales. Each scale ranges from 0 to 100, with higher scores indicating better physical functioning, role functioning, or GHS and QoL, respectively.⁴² EORTC QLQ-C30 was administered to adult patients only in both the COMMODORE 2 and COMMODORE 1 studies.

EORTC Item Library: PNH Symptoms

The EORTC Item Library is a database of items used in fully and partially validated EORTC QoL questionnaires. Eight items from the library were used to assess disease-related symptoms that are relevant to patients with PNH that are not sufficiently covered in the other PRO measures (EORTC IL-40). Selected symptoms included headache, dyspnea at rest and upon exertion, dysphagia, abdominal pain, chest pain, and erectile dysfunction.^39,42,46 All items are scored on a 4-point scale that ranges from “not at all” to “very much,” with higher scores indicative of higher symptom severity. All symptom items will be scored individually unless they comprise a scale (i.e., dyspnea). This measure was administered to adult patients only in both trials.

Patient Preference Questionnaire

The Patient Preference Questionnaire (PPQ) is a 2-item measure of treatment preference developed by the sponsor. Patients were asked to indicate their preference for IV eculizumab or ravulizumab (the COMMODORE 1 study’s arm C only) or SC crovalimab and indicate the top 3 reasons for their preference from a list of 13 potential reasons.^33,34 In the COMMODORE 2 study, this measure was administered to patients aged 18 years or older in arm B. In the COMMODORE 1 study, this measure was administered to patients aged 12 years or older at week 17 in arm A and arm C, and administered at week 41 to all 3 arms.

Pediatric Quality of Life Inventory Multidimensional Fatigue Scale

The Pediatric Quality of Life Inventory (PedsQL) Multidimensional Fatigue Scale (MFS) is a valid, reliable measure for assessing fatigue in healthy children and adolescents, and those with a range of acute and chronic health conditions.^47-50 Self-report versions are available for children and adolescents that contain developmentally appropriate language for defined age groups (i.e., 5 years to 7 years, 8 years to 12 years, and 13 years to 17 years). This version contains 18 items that are scored in 3 respective domains: general fatigue (6 items), sleep/rest fatigue (6 items), and cognitive fatigue (6 items). Domains can be further combined into an overall total fatigue score. Domain and total scores are converted to a 0 to 100 scale, with higher scores indicative of lower fatigue. For the COMMODORE 2 and COMMODORE 1 studies, the self-report version with a 1-week recall period (i.e., acute version) was used for patients aged 8 years to 17 years.

PedsQL Generic Core Scales (Version 4)

The PedsQL Generic Core Scales are valid, reliable measures for assessing HRQoL in healthy children and adolescents, and those with acute or chronic health conditions.^51,52 Self-report versions are available for children and adolescents that contain developmentally appropriate language for defined age groups (i.e., 5 years to 7 years, 8 years to 12 years, and 13 years to 17 years). This version contains 23 items that are scored in 4 respective domains: physical functioning (8 items), emotional functioning (5 items), social functioning (5 items), and school functioning (5 items). Physical and psychosocial health summary scores, as well as an overall total score, can be created by further combining domains. Domain and total scores are converted to a 0 to 100 scale, with higher scores indicative of better functioning. Only the physical functioning scale of the self-report version with a 1-week recall period (i.e., acute version) was used for patients aged 8 years to 17 years.

Harms Outcomes

In the COMMODORE 2 and COMMODORE 1 studies, safety was assessed through descriptive summaries of exposure to each study treatment, AEs, changes in laboratory test results, and changes in vital signs and an electrocardiogram (ECG). Safety analyses were performed on the safety-evaluable population, defined as all enrolled participants who received at least 1 dose of study drug (crovalimab, eculizumab, or placebo), with patients grouped according to treatment received.

Considerations that informed the selection of efficacy and safety outcomes to be summarized and assessed using GRADE included the following:

• the proportion of patients with hemolysis control was identified as important by the patient and clinician group input, and by the clinical experts consulted for this review

• the proportion of patients with BTH was identified as important by the clinical experts

• the mean percentage change in LDH levels was identified as important by the clinical experts

• the proportion of patients with stabilized hemoglobin was identified as important by the clinical experts

• the proportion of patients who reached and maintained a minimum hemoglobin level was identified as important by the clinical experts

• the proportion of patients who achieved transfusion avoidance was identified as important by the clinical experts

• the mean number of total pRBC units transfused was identified as important by the clinical experts

• HRQoL outcomes were identified by the patient input, and the clinical experts and clinician group input specified to include the FACIT-F and the EORTC QLQ-C30 GHS and QoL measures

• the proportion of patients with infections (including meningococcal meningitis) and the proportion of deaths were identified as important safety outcomes by referring to other relevant reviews.

The following outcomes were considered relevant in providing supplementary data in assessing the efficacy and safety of crovalimab for PNH by the clinical experts, but only needed to be summarized in the report without using the GRADE tool for assessment:

• the proportion of patients who achieved a 5-point or greater improvement from baseline in the FACIT-F score at week 5

• the mean change in EORTC IL-40 symptoms for 6 subscales (dyspnea, dysphagia, chest pain, headaches, abdominal pain, and erectile dysfunction)

• the proportion of patients with a preference for crovalimab or eculizumab

• the proportion of patients with a MAVE

• the proportion of patients with an immune complex–mediated reaction (only applicable among the patients who switched from other C5 inhibitors to crovalimab).

Table 8: Summary of Outcome Measures and Their Measurement Properties

ARC = absolute reticulocyte count; EORTC = European Organisation for Research and Treatment of Cancer; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; GHS = global health status; Hb = hemoglobin; HRQoL = health-related quality of life; MID = minimal important difference; PNH = paroxysmal nocturnal hemoglobinuria; QoL = quality of life.

Statistical Analysis

The statistical analysis of efficacy end points in the COMMODORE 2 and COMMODORE 1 trials is summarized in Table 9.

Sample Size and Power Calculation

COMMODORE 2 Study

The sample size estimation for the randomized portion of the study (arm A and arm B) was based on the noninferiority assessment of the coprimary end points of hemolysis control, as assessed by centrally measured LDH levels (the assumption of the rate of hemolysis control was 86% in both the crovalimab and the eculizumab arms), and the proportion of patients who achieve transfusion avoidance (the assumption of the rate of transfusion avoidance was 67% in the crovalimab arm and 66% in the eculizumab arm) during the efficacy period. The final target sample size corresponds to the end point that requires the larger number of patients (i.e., transfusion avoidance from baseline to week 25). A total of 204 adult patients were randomly assigned in a 2:1 ratio to receive either crovalimab (n = 135) or eculizumab (n = 69) to ensure approximately 180 evaluable patients, assuming a 10% dropout rate. This sample size provided 80% power to demonstrate the noninferiority of crovalimab to eculizumab with respect to transfusion avoidance, using a NIM of –20% and 1-sided type I error rate of 2.5%.¹⁵

The NIM for transfusion avoidance was determined based on the data (protocol of ALXN1210-PNH-301 [the 301 study])⁶⁰ comparing eculizumab-treated patients with untreated patients from the Global PNH Patient Registry for eculizumab-treated patients (i.e., patients treated with eculizumab showed a benefit over untreated patients), with a difference of approximately 40% (with a transfusion avoidance proportion of 57.1% and 18.6%, respectively), after adjustment for the history of transfusions 12 months before enrolment. It was assumed that a difference in proportions of –20%, the NIM, would preserve at least 50% of the control treatment effect. This NIM was also defined considering the rarity of PNH. A more conservative NIM would have resulted in the estimated sample size being too large and infeasible. Lee et al. reported a proportion of patients with transfusion avoidance of 66.1% (95% CI, 65.9% to 74.6%) in treatment-naive patients treated with eculizumab.⁶⁰

With regard to hemolysis control, 116 patients were required in a 2:1 ratio to test the noninferiority of crovalimab versus eculizumab, with a NIM of 0.2 in the OR scale, 80% power, and 1-sided type I error rate of 2.5%. Incidentally, a similar sample size was required to test for noninferiority in the probability scale when the NIM was –0.2. The 301 study also reported a proportion of LDH normalization below 1.0 multiplied by ULN of 49.4%.⁶⁰ Under the assumption of LDH being distributed via a log-normal method, the expected proportion below 1.5 multiplied by ULN was 86%. The same proportion was assumed for crovalimab. The NIM in the OR scale was obtained as 1 divided by OR,^0.5 where OR equals 24.6 assuming 86% of patients receiving eculizumab will reach an LDH level of 1.5 or less multiplied by ULN compared to an upper bound of the 95% CI of the proportion among placebo-treated patients of 20%. Note that both proportions are approximately twice the ones used in the 301 study⁶⁰ for an LDH level of 1 or less multiplied by ULN, and a fraction of 0.5 of that effect is retained.⁶¹ Assuming a 10% dropout, the total needed sample size would be 128 patients (85 patients randomized to crovalimab and 43 patients to eculizumab). With 180 evaluable patients expected in the study, the power for this end point would be 94%. Hence, the joint power for both transfusion avoidance and LDH would be 75% if they were uncorrelated. It is likely that coprimary end points would be correlated and hence, joint power would be higher. If there were no dropouts and all the approximately 200 patients (the calculated sample size) contributed at least 1 LDH sample, then the power for transfusion avoidance would be 84% and for LDH 96%.

Pediatric patients were enrolled in the descriptive arm (arm C) throughout the duration of the study. No target sample size was specified.

The primary analysis was performed once the last randomized patient on trial completed 24 weeks of study treatment or discontinued early, whichever happened first.¹⁵

COMMODORE 1 Study

The COMMODORE 1 study was originally planned to recruit 200 patients to the randomized arms to be sufficiently powered to assess the efficacy of crovalimab. Considering the evolving treatment landscape and a reduced pool of patients treated with eculizumab over time, key design features of the study had to be modified in the protocol amendment version 6. As per protocol amendment version 6, the safety objective was the only primary objective of the COMMODORE 1 study and all efficacy analyses were considered descriptive without any formal statistical testing. In the modified design, the randomized arms were expected to randomize approximately 90 adult patients (aged ≥ 18 years) in a 1:1 ratio to crovalimab (arm A) or eculizumab (arm B). The nonrandomized arm C originally intended to recruit about 50 patients in different clinically relevant cohorts. With protocol amendment version 6, arm C was also further expanded to include an additional cohort of adult patients (≥ 18 years) who had been receiving eculizumab at the approved dose for at least 24 weeks before study entry (i.e., the same population as those who would have been enrolled in the randomized arms, to continue study access for such patients in a nonrandomized setting). With this, the total sample size of arm C was increased to approximately 100 patients.¹⁷

In line with protocol amendment version 6, the primary analysis for this study was conducted at the same time as the primary analysis of the COMMODORE 2 study, which was the pivotal trial assessing crovalimab in naive patients with PNH. Patients had to have received at least 1 dose of crovalimab or eculizumab to be included in the primary analysis of the COMMODORE 1 study.¹⁷

Statistical Testing

COMMODORE 2 Study

Coprimary end point: Transfusion avoidance — The percentage of patients with transfusion avoidance from baseline through week 25 (after 24 weeks on treatment) was computed. As a conservative approach, patients who prematurely withdrew from the study before week 25 were assumed to have undergone a transfusion. The difference in the percentage of patients with transfusion avoidance in the 2 treatment arms was calculated, along with a 95% CI for the difference using the stratified Newcombe CI method.⁶² The difference between the 2 treatment arms was computed as a weighted combination of the differences between the crovalimab and eculizumab arms within the stratification indicators of transfusion history and the LDH level before randomization using Mantel-Haenszel weights.⁶³ Noninferiority with respect to transfusion avoidance was to be concluded if the lower limit of the 95% CI for the difference between crovalimab and eculizumab was greater than the predefined NIM of −20% (alternative hypothesis).

Coprimary end point: Hemolysis control — A generalized estimating equation (GEE) model was used to estimate the adjusted log OR of hemolysis control (LDH level ≤ 1.5 × ULN) due to treatment, accounting for the intraindividual correlation between LDH control statuses across visits. Covariates included the categorical effects of treatment and visit, visit by treatment interaction, continuous baseline LDH level, and number of pRBC units administered within the 6 months before randomization. The baseline LDH level was defined as the mean of all central LDH values taken during screening, and the LDH value at week 1, day 1, collected before the first dose administration of either crovalimab or eculizumab. The primary analysis applied an unstructured covariance matrix, which imposed minimal assumptions; however, it required estimating a large number of parameters, which may prevent model convergence. The sensitivity analysis accounted for this (described as follows in the sensitivity analyses section).

Measurements taken every 2 weeks from week 5 through week 25 corresponded to the period of stability in treatment-naive patients and were used in the statistical analysis, given that treatment-naive patients were expected to reach a stable LDH plateau by the end of the first month of treatment. For each patient at each visit, a binary variable was created if the central LDH level was less than or equal to 1.5 times the ULN or if the central LDH level was more than 1.5 times the ULN, and it was considered missing if LDH was not measured. Unscheduled central LDH measurements were mapped to the nearest scheduled assessment visit using time windowing. If more than 1 measurement was attributed to a visit window, the average was taken. Central LDH samples affected by tabletop hemolysis were excluded from analyses. The noninferiority of crovalimab relative to eculizumab was concluded if the lower limit of the 2-sided 95% CI for the OR of crovalimab compared to eculizumab was greater than the predefined NIM of 0.2 (alterative hypothesis).

Secondary end points: If noninferiority was established for the coprimary end points, superiority testing of primary and secondary end points was conducted using the hierarchical order as described in Table 9 to ensure that the familywise 1-sided type I error rate was controlled at the 2.5% level. A closed testing procedure was used and the lack of significance of an end point in the hierarchy precluded the analysis of subsequent end points.

Table 9: Hierarchical Statistical Testing of Primary and Secondary Efficacy End Points in the COMMODORE 2 Study

BTH = breakthrough hemolysis; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; TA = transfusion avoidance.

^aDenotes coprimary efficacy end points that were tested together.

Source: Sponsor’s submission.²³

BTH: The proportion of patients with BTH from baseline through week 25 was analyzed using the same methodology as for transfusion avoidance. As a conservative approach, patients withdrawing from study treatment before week 25 were assumed to have experienced a BTH event in the unobserved period. If the upper limit of the 95% CI for the difference between crovalimab and eculizumab in the proportion of patients with BTH was less than the predefined NIM of 20%, then crovalimab was considered noninferior to eculizumab.

Hemoglobin stabilization: The stabilization of hemoglobin was analyzed using the same approach as for transfusion avoidance. As a conservative approach, patients who withdrew from study treatment before week 25 were assumed to not have met hemoglobin stabilization criteria. If the lower limit of the 95% CI for the difference between crovalimab and eculizumab in the proportion of patients with stabilized hemoglobin was greater than the predefined NIM of −20%, then crovalimab was to be considered noninferior to eculizumab.

FACIT-F: The change from baseline to week 25 was analyzed using a mixed model for repeated measures (MMRM) assuming normally distributed scores, with adjustment for stratification factors, and the baseline FACIT-F score. An unstructured covariance matrix was used to model the within-patient errors. If the model did not converge with the unstructured covariance matrix, a more parsimonious structure was considered in the following order until model convergence was achieved: Toeplitz, first-order autoregressive, and compound symmetry. If the lower limit of the 95% CI for the difference between crovalimab and eculizumab in the mean change from baseline to week 25 in fatigue was greater than the predefined NIM of −5 points, then crovalimab was considered noninferior to eculizumab and the end point was to be tested for superiority (Table 9).

Exploratory efficacy analysis: All exploratory efficacy end points were analyzed descriptively, without any formal statistical testing. No adjustment for multiplicity was conducted. Patient preference was assessed in arm B patients at switch week 17 (Week 41). The intermediate time point was chosen to allow patients sufficient time on treatment with crovalimab to be able to compare to eculizumab while preventing recall bias.²³

COMMODORE 1 Study

As the anticipated enrolment was not expected to provide sufficient power for the statistical testing of efficacy data from the randomized arms, all efficacy objectives and related end points became exploratory and descriptive. There was no change to the safety end points and the safety analysis was conducted as per the COMMODORE 2 study.²³

For all exploratory efficacy end points, summary statistics (including means, median, range, SDs, and proportions where appropriate) were presented by treatment arm using tables, listings, and graphs as appropriate. Proportions were reported with 95% CIs. In graphs, CIs were displayed for a treatment arm when the sample size was at least 10.

In nonrandomized arm C, the same analyses as for the randomized arms are repeated, except for the LDH level percentage change. Change in LDH levels from baseline to LDH levels averaged over week 21, week 23, and week 25 were computed. In contrast to patients in arm A and arm B, not all patients in arm C were required to have hemolysis control (defined as an LDH level ≤ 1.5 × ULN) at study entry, and therefore the mean change in the LDH level from baseline is not a meaningful analysis for arm C.

Exploratory Efficacy End Points for Randomized Arm A and Arm B

• Percentage change from baseline in LDH levels:

  • Baseline LDH was defined as the mean of all central LDH values taken during screening (i.e., within 4 weeks before the first on-study drug administration of either crovalimab or eculizumab), and the LDH value at week 1, day 1, collected before the first dose administration of either crovalimab or eculizumab. All postbaseline LDH values included in this analysis were based on the central laboratory LDH measurements. All LDH measurements (up to week 25) were used in the statistical model. Unscheduled central LDH measurements were mapped to the nearest scheduled assessment visit using time windowing. If more than 1 measurement was attributed to a visit window, the average was taken. Central LDH samples affected by tabletop hemolysis were excluded from analyses. One additional LDH sample was taken at week 1, day 2, only for patients on the crovalimab arm following the week 1, day 2, crovalimab loading dose. This assessment was not included in the statistical model because there was no corresponding LDH sample (week 1, day 2) for patients in the eculizumab arm.

  • The percentage change in LDH levels from baseline averaged over week 21, week 23, and week 25 between the crovalimab arm (arm A) and the eculizumab arm (arm B) were analyzed using the MMRM with fixed categorical effects of treatment, visit, and visit by treatment arm interaction, as well as the continuous fixed covariate of baseline LDH level, baseline LDH level by visit interaction, and the stratification factor of transfusion history (yes or no).⁶⁴ An unstructured covariance matrix was used to model the within-patient errors. If the model did not converge with the unstructured covariance matrix, a more parsimonious structure was considered in the following order until model convergence was achieved: Toeplitz, first-order autoregressive, and compound symmetry. The point estimate for the difference between crovalimab and eculizumab, along with a 2-sided 95% CI, was presented. The percentage change in LDH levels from baseline averaged over week 21, week 23, and week 25 was estimated using contrasts. The model included data from all patients in the efficacy-evaluable population. Missing data were assumed as data missing at random as part of the MMRM model.^17,23

• Hemolysis control:

  • A GEE model was used to estimate the adjusted log OR of an LDH level of 1.5 or less multiplied by ULN due to treatment, taking account of the intraindividual correlation between LDH control statuses across visits. The GEE model was adjusted for baseline covariates and correlation matrix structure in a manner similar to the MMRM. Similarly, unscheduled LDH assessments were included in the analysis via windowing, and LDH samples suspected of being affected by tabletop hemolysis were excluded from analyses.

• Transfusion avoidance:

  • The percentage of patients with transfusion avoidance from baseline through week 25 (after 24 weeks on treatment) was computed. Note that, as a conservative approach, patients who prematurely withdraw from study treatment were assumed to have undergone a transfusion. For the descriptive comparison of transfusion avoidance between the randomized arms, a difference in the percentage of patients with transfusion avoidance in the 2 treatment arms was calculated, along with a 95% CI for the difference using the stratified Newcombe CI method. ⁶² The difference between the 2 treatment arms was computed as a weighted combination of the differences between the crovalimab and eculizumab arms within the stratification indicator of transfusion history using Mantel-Haenszel weights. ⁶³ The total number of units (based on a local equivalent) of pRBCs transfused per patient by week 25 was presented with 95% CI per treatment arm. Transfusion avoidance analysis included only patients who had completed the 24 weeks of treatment, or who had discontinued from study but had their first dose of the relevant study treatment at least 24 weeks before CCOD.

• Breakthrough hemolysis:

  • The proportion of patients with BTH from baseline through week 25 was analyzed using the same approach as for transfusion avoidance. As a conservative approach, patients withdrawing from study treatment before week 25 were assumed to have experienced BTH in the unobserved period.

• Hemoglobin stabilization and other analyses of proportion:

  • The proportion of patients with stabilization of hemoglobin from baseline through week 25 was analyzed using the same approach as for transfusion avoidance. As a conservative approach, patients withdrawing from study treatment before week 25 were assumed to not have met hemoglobin stabilization criteria. Similar analyses of proportions were performed for other end points (e.g., the proportion of patients who had experienced a MAVE).

• Analysis of PRO end points:

  • Changes from baseline in HRQoL end points were calculated at each visit. Summary statistics describe the raw scores and mean change (95% CI) from baseline at each visit. Plots of average scores with dispersion bars were presented over time and by treatment arm.

  • The FACIT-F score was further analyzed as change from baseline to week 25 using the MMRM approach.

  • For patient preference for crovalimab, or eculizumab or ravulizumab, the percentage of patients preferring each treatment and the reasons for preference were summarized per treatment arm.

Sensitivity Analyses

COMMODORE 2 Study

Various sensitivity analyses were performed to assess the robustness of the coprimary end point results against the following.

• Efficacy population definition: Hemolysis control and transfusion avoidance were assessed in the following alternative efficacy populations —

  • randomized (ITT) population

  • per-protocol population.

• Impact of missing data: The following multiple imputation approaches were performed to assess the robustness of the GEE model against various missing data assumptions in the analysis of hemolysis control —

  • A Markov Chain Monte Carlo (MCMC) model was used assuming missing data are missing at random.

  • Pattern mixture models and tipping point analyses were used assuming data are missing not at random. Given that the pattern mixture model imputes missing observations to follow the distribution of the control arm, the approach is considered anticonservative. The tipping point analysis was performed by first assuming a worst case scenario (LDH level ≤ 1.5 × ULN not met for all missing central LDH values) in the crovalimab arm and the reverse for the eculizumab arm (LDH level ≤ 1.5 × ULN met for all missing central LDH values). Only if this worst case scenario was not consistent with the primary analysis would intermediary imputations be performed.

  • It was expected that all transfusions were recorded.

COMMODORE 1 Study

The sensitivity analyses in the COMMODORE 1 study were not reported.

Subgroup Analyses

COMMODORE 2 Study

The robustness of the treatment effect of crovalimab compared to eculizumab — assessed via the coprimary efficacy end points of hemolysis control and transfusion avoidance — was investigated in predefined subgroups based on key baseline demographic and disease characteristics (age, sex, region [Europe/Central, South, and North America; or Japan and the rest of Asia-Pacific], eculizumab available in region [yes or no], race, pRBC units transfused in the 6 months before baseline, local LDH level at randomization, body weight, and prior diagnosis of aplastic anemia).^15,35

Table 10: Statistical Analysis of Efficacy End Points

AE = adverse event; BTH = breakthrough hemolysis; CI = confidence interval; ECG = electrocardiogram; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; GEE = generalized estimating equation; ITT = intention to treat; LDH = lactate dehydrogenase; MCMC = Markov Chain Monte Carlo; MMRM = mixed model of repeated measures; NA = not applicable; PMM = predictive mean matching; pRBC = packed red blood cell; TA = transfusion avoidance; ULN = upper limit of normal.

Sources: COMMODORE 2 Primary Clinical Study Report,¹⁵ COMMODORE 1 Primary Clinical Study Report,¹⁷ and sponsor’s submission.¹⁶

COMMODORE 1 Study

As the study evolved into a safety study, subgroup analyses for efficacy were not considered or summarized in this review.

Analysis Populations

The definitions for the different analysis populations in the COMMODORE 2 and COMMODORE 1 trials are summarized in Table 11.

Table 11: Analysis Populations of the COMMODORE 2 and COMMODORE 1 Studies

CCOD = clinical cut-off date; ITT = intention to treat; LDH = lactate dehydrogenase; PAP = primary analysis population; PNH = paroxysmal nocturnal hemoglobinuria; PP = per protocol; SNP = single nucleotide polymorphism; ULN = upper limit of normal; WBC = white blood cell.

^aFor analysis of end points defined over 24-week treatment intervals, only patients with at least 24 weeks of crovalimab treatment were included. However, listings and by visit summary analysis of efficacy end points included all available efficacy data.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and COMMODORE 1 Primary Clinical Study Report.¹⁷

Results

Patient Disposition

Patient disposition for the COMMODORE 2 and COMMODORE 1 trials is summarized in Table 12.

COMMODORE 2 Study

Of the 239 patients in the randomized population screened in the COMMODORE 2 study, 29 (12%) patients were those who did not advance past screening, leading to an enrolment of 210 patients in total. The main reasons for not advancing past screening were poor willingness to comply (62%) and absolute neutrophil count of less than 500 cells/µL (14%). A total of 204 patients were randomized. Of the 135 patients randomized to crovalimab (arm A), 129 (95.6%) patients completed 24 weeks of treatment in the primary treatment period and continued to receive crovalimab in the crovalimab extension period. Of these, 127 patients continued to receive crovalimab treatment up to the CCOD. Of the 69 patients randomized to eculizumab (arm B), 68 (98.6%) patients completed 24 weeks of treatment in the primary treatment period and switched to crovalimab treatment in the crovalimab extension period. Of these 68 patients, 65 continued to receive crovalimab treatment up to the CCOD; the other 3 patients discontinued study treatment and did not enter the safety follow-up. All patients in the randomized arms were 18 years or older, with the exception of 2 adolescent patients (both aged 17 years at the time of randomization), who were randomized to the eculizumab arm before the opening of the separate pediatric-specific descriptive arm C in protocol version 3, and switched to crovalimab in the crovalimab extension period after completing the primary treatment period. One of the 2 adolescent patients had already turned 18 years at the time of switching to crovalimab in the extension period while the other patient was still aged 17 years at the time of switching.¹⁵

Table 12: Summary of Patient Disposition From the COMMODORE 2 and COMMODORE 1 Studies

ANC = absolute neutrophil count; CCOD = clinical cut-off date; ITT = intention to treat; LDH = lactate dehydrogenase; NA = not applicable; NR = not reported; PP = per protocol.

Note: Data presented in Table 12 were based on analyses at the CCOD of November 16, 2022.

^aTwo pediatric patients were enrolled in randomized arm B before separate descriptive arm C was opened in protocol version 3.

^bThree patients did not enter the safety follow-up.

^cPatient did not enter the safety follow-up.

^dDeath due to myocardial infarction at day 2, which was unrelated to the study drug.

^eDeath due to fatal ischemic stroke at day 71, which was unrelated to the study drug.

^fSix patients were ongoing in the arm A primary treatment period at CCOD.

^gFive patients were ongoing in the arm B primary treatment period at CCOD.

^hFive patients were ongoing in the arm C primary treatment period at CCOD.

ⁱDeath due to colorectal cancer, which was unrelated to the study drug.

^jThe patient who died in arm A (crovalimab) was excluded from the primary analysis population for not fulfilling the criteria of at least 1 postbaseline LDH level assessment.

^kArm B switch patients were defined as all patients in arm B who received at least 1 dose of crovalimab from week 25, and had at least 1 valid postswitch LDH level assessment by the central laboratory after the first IV infusion.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and COMMODORE 1 Primary Clinical Study Report.¹⁷

COMMODORE 1 Study

Of the 146 patients in the randomized population screened in the COMMODORE 1 study, 19 (13%) patients did not advance past screening due to not meeting the inclusion criteria, leading to an enrolment of 127 patients in total. The main reasons for not advancing past screening were poor willingness to comply (32%) and documented eculizumab treatment as per label (21%). In the nonrandomized arm C, 38 patients were enrolled at the time of the CCOD. Of the 45 patients randomized to the crovalimab arm, 39 (86.7%) patients completed 24 weeks of treatment and then continued to receive crovalimab treatment in the crovalimab extension period. Of these, 37 patients continued to receive crovalimab treatment up to the CCOD. Of the 45 patients in the crovalimab arm, 5 patients were still ongoing in the primary treatment period as of the CCOD. Of the 44 patients randomized to the eculizumab arm, 35 (79.5%) patients completed 24 weeks of eculizumab treatment and switched to crovalimab treatment upon entering the crovalimab extension period. Of these, 32 patients continued to receive crovalimab treatment up to the CCOD. Of the 44 patients in the eculizumab arm, 5 patients were still ongoing in the primary treatment period as of the CCOD.¹⁷

Baseline Characteristics

The baseline characteristics outlined in Table 13 are limited to those that are most relevant to this review or were relevant for the interpretation of the study results.

COMMODORE 2 Study

In the COMMODORE 2 study, the randomized population had a median age of 36 years (crovalimab arm) or 38 years (eculizumab arm), and most patients were within the age range of 18 years to 64 years (90% in the crovalimab arm and 84% in the eculizumab arm).¹⁵ Approximately 43% to 49% were females and 51% to 57% of the patients were males. The majority of patients were Asian (64% to 74%) or white (23% to 33%), followed by other or unknown (3%). Most patients weighed between 40 kg and less than 100 kg at baseline (97% in the crovalimab arm versus 96% in the eculizumab). The baseline mean LDH level was 7.6 multiplied by ULN and 7.8 multiplied by ULN in the crovalimab and eculizumab arms, respectively. Most patients had an LDH level of more than 4 multiplied by ULN (83% in the crovalimab arm versus 84% in the eculizumab arm). At baseline, 74% to 77% of the patients had received a pRBC transfusion with mean transfused pRBCs within the 12 months before screening of 6.5 (SD = 8.3) units in the crovalimab arm and 6.6 (SD = 8.7) units in the eculizumab arm, with half of the patients receiving up to 6 units of pRBC transfusion in the prior 6 months in both arms. The median of the PNH clone size was smaller in the crovalimab arm than in the eculizumab arm for erythrocytes (25% and 45%, respectively) and granulocytes (60% versus 75%, respectively). The main PNH-relevant conditions in patient history were aplastic anemia (39% in the crovalimab arm and 38% in the eculizumab arm), major vascular events (16% versus 15%, respectively), and MDS (4% versus 9%, respectively).

Table 13: Summary of Baseline Characteristics From the COMMODORE 2 and COMMODORE 1 Studies

ECU = eculizumab; LDH = lactate dehydrogenase; MDS = myelodysplastic syndrome; NA = not applicable; NE = not estimable; NR = not reported; PNH = paroxysmal nocturnal hemoglobinuria; pRBC = packed red blood cell; RAVU = ravulizumab; SD = standard deviation; SNP = single nucleotide polymorphism; ULN = upper limit of normal.

Note: Data presented in Table 13 were based on analyses at the clinical cut-off date of November 16, 2022. Only patients with available data were included in the calculations of percentages.

^aTwo patients aged 17 years were randomized to the ECU arm before the opening of the separate pediatric-specific nonrandomized arm.

^bBaseline LDH level was defined as the mean of all central LDH values, collected within 28 days before the first on-study drug administration including the predose value from day 1. The switch baseline LDH value was defined as the mean of all central LDH values, collected within 28 days of the first dose of crovalimab including the predose value from switch day 1.

^cThe maximum baseline hemoglobin value in the ECU arm of 810 g/L was a result of erroneous data entry.

^dNo patients in this cohort had data available for the PNH clone size (%) for granulocytes or monocytes.

Sources: COMMODORE 2 Primary Clinical Study Report,¹⁵ COMMODORE 1 Primary Clinical Study Report,¹⁷ and sponsor’s submission.²³

COMMODORE 1 Study

In the COMMODORE 1 study, the randomized population had a median age of 42 years (crovalimab arm) or 49 years (eculizumab arm), and as in the COMMODORE 2 study, most patients were within the age range of 18 years to 64 years (89% in the crovalimab arm and 84% in the eculizumab arm). Approximately half of the patients were female (50% to 53%) and half were male (47% to 50%). Unlike the COMMODORE 2 trial, the majority of patients in the COMMODORE 1 trial were white (76% in the crovalimab arm and 73% in the eculizumab arm), followed by Asian (20% versus 16%, respectively), and other or unknown (4% versus 11%, respectively). Most patients weighed between 40 kg and less than 100 kg (93% in the crovalimab arm versus 91% in the eculizumab arm). The mean time from PNH diagnosis to enrolment was shorter in the crovalimab arm (8.0 [SD = 6.6] years) than in the eculizumab arm (11.2 [SD = 7.1] years). At baseline, 23% to 25% of the patients had received a pRBC transfusion with mean pRBC units transfused within the 12 months before screening of 1.6 (SD = 3.7) units in the crovalimab arm and 2.3 (5.4) units in the eculizumab arm. The median of PNH clone size for monocytes was larger in the eculizumab arm (96%) than in the crovalimab arm (89%). The main PNH-relevant conditions in patient history were aplastic anemia (33% in the crovalimab arm and 36% in the eculizumab arm) and major vascular events (22% versus 23%, respectively).

Exposure to Study Treatments

The extent of patient exposure to the study drugs in the COMMODORE 2 and COMMODORE 1 trials is summarized in Table 14.

COMMODORE 2 Study

The median treatment duration during the primary treatment period was similar in the crovalimab and eculizumab arms (20.1 weeks versus 22.1 weeks, respectively). The median number of doses administered in the crovalimab arm at 10 (range, 1 to 12) was lower than in the eculizumab arm at 14 (range, 6 to 16). The median total cumulative dose in the crovalimab arm at 5,760 mg (range, 1,000 mg to 9,390 mg) was lower than in the eculizumab arm at 11,400 mg (range, 4,200 mg to 13,200 mg).

In the crovalimab arm up to the CCOD (November 16, 2022), the median treatment duration was 48.3 weeks, with 57% of patients having had a treatment duration of at least 48 weeks. Analysis of serum concentrations showed that 5 patients had partial loss of crovalimab exposure and 2 patients had complete loss of exposure.¹⁵

At the CCOD, the median treatment duration with crovalimab for eculizumab (arm B) switch patients was 24.1 weeks, with a median of 11 doses administered. The median total cumulative dose was 6,440 mg (range, 1,340 mg to 15,280 mg). Overall, 40 (59%) patients had a treatment duration of at least 24 weeks and 41% of patients had a treatment duration of less than 24 weeks. One patient reached more than 72 weeks of treatment duration at the CCOD.¹⁵

In arm C, treatment duration in the 6 pediatric patients ranged between 21 weeks and 57 weeks. All patients completed the 24-week primary treatment period per the dose recommendation in the study protocol and were ongoing in the crovalimab extension period as of the CCOD. For 1 pediatric patient, the SC dose of crovalimab was increased from 680 mg to 1,020 mg from week 13 onward in the context of weight increase to more than 100 kg (from 98.5 kg at baseline); however, this weight increase did not meet the more than 10% change criterion and a major protocol deviation in relation to this modification was recorded; no new safety signal was detected in relation to this increased crovalimab dose. The patient remained on the 1,020 mg dose at the time of the CCOD.¹⁵

COMMODORE 1 Study

The median treatment duration during the primary safety period was 20.1 weeks in the crovalimab arm and 22.1 weeks in the eculizumab arm. The median number of doses administered was 10 (range, 4 to 11) and 12 (range, 1 to 14) in the crovalimab and eculizumab arms, respectively. The median total cumulative dose in the crovalimab arm at 5,760 mg (range, 2,020 mg to 7,960 mg) was lower than in the eculizumab arm at 10,800 mg (range, 900 mg to 12,600 mg).

In the crovalimab arm up to the CCOD, the median treatment duration was 52 weeks. The majority of patients (54.5%) had a treatment duration of at least 48 weeks. At the CCOD, the median treatment duration with crovalimab for eculizumab (arm B) switch patients was 32 weeks, with a median of 13 doses administered. The median total cumulative dose was 9,160 mg (range, 2,360 mg to 17,140 mg). The majority of patients (74%) had a treatment duration of at least 24 weeks and 26% of patients had a treatment duration of less than 24 weeks.¹⁷

In arm C, the median treatment durations in the crovalimab safety population were 2.1 weeks (range, 2.1 weeks to 2.1 weeks), 36.0 weeks (range, 2.3 weeks to 80.1 weeks), 38.1 weeks (range, 0.3 weeks to 64.1 weeks), and 50.2 weeks (range, 28.1 weeks to 100.1 weeks) in the pediatric, prior ravulizumab, prior high-dose eculizumab, and C5 polymorphism cohorts, respectively. One patient in the pediatric cohort, 6 patients in the prior ravulizumab cohort, and 2 patients in the prior high-dose eculizumab cohort, respectively, had a treatment duration of less than 24 weeks at the time of primary analysis. The median number of doses administered was 4.0 (range, 4.0 to 4.0), 14.0 (range, 4.0 to 25.0), 14.5 (range, 2.0 to 20.0), and 17.5 (range, 12.0 to 30.0) in the pediatric, prior ravulizumab, prior high-dose eculizumab, and C5 polymorphism cohorts, respectively. The median total cumulative doses were 2,020 mg (range, 2,020 mg to 2,020 mg), 8,480 mg (range, 2,020 mg to 15,960 mg), 8,820 mg (range, 1,340 mg to 12,255 mg), and 10,860 mg (range, 7,120 mg to 19,360 mg) in the pediatric, prior ravulizumab, prior high-dose eculizumab, and C5 polymorphism cohorts, respectively. The median IV and SC dose intensity across these cohorts was 100.0% (IV range, 100.0% to 137.5%; SC range, 85.7% to 102.4%).¹⁷

Table 14: Summary of Patient Exposure From the COMMODORE 2 and COMMODORE 1 Studies (Primary Safety Period, Randomized Safety Population)

IQR = interquartile range; NA = not applicable; NE = not estimable; SC = subcutaneous; SD = standard deviation.

Note: Data presented in Table 14 were based on analyses at the clinical cut-off date of November 16, 2022. IV dose intensity is a percentage based on actual IV dose per total planned IV dose. SC dose intensity is a percentage based on actual SC dose per total planned SC dose.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and COMMODORE 1 Primary Clinical Study Report.¹⁷

Adherence to Self-Administration in COMMODORE 2 and COMMODORE 1 Studies

Crovalimab self-administration or drug administration by a caregiver was permitted in both the COMMODORE 2 and COMMODORE 1 studies starting at week 9, after training and confirmation of proficiency by the health care provider. Between week 9 and week 25, the majority of SC injections were administered by health care professionals with an increasing proportion of patients with SC administration of crovalimab done by the patient or their caregiver up to week 25 (range, 17.1% to 31.0%, in arm A of the COMMODORE 2 study; range, 11.6% to 15.8%, in arm A of the COMMODORE 1 study; and range, 21.6% to 27.6%, in arm C of the COMMODORE 1 study). After week 25, the rate of administration by the patient or caregiver was generally rising, with an increase in offsite administrations. No medication errors specifically due to self-administration were reported up to the CCOD in either study.¹⁵

Concomitant Medications and Cointerventions in COMMODORE 2 and COMMODORE 1 Studies

The most frequently used (≥ 50% in either arm) previous or concomitant medications in the 2 trials are listed in Table 15.

In the COMMODORE 2 study, the majority of patients in the crovalimab and eculizumab arms had at least 1 previous medical condition (74.1% versus 81.2%, respectively).¹⁵ The most frequent previous medical conditions were hypertension (22.2% in the crovalimab arm versus 14.5% in the eculizumab arm), cholelithiasis (6.7% versus 8.7%, respectively), and MDS (4.4% versus 8.7%, respectively).¹⁵ Most patients in each arm received at least 1 previous or concomitant treatment (99.3% in the crovalimab arm versus 98.6% in the eculizumab arm). The most common concomitant medications were ophthalmologicals (78.5% in the crovalimab arm versus 87.5% in the eculizumab arm), antibacterials for systemic use (76.3% versus 75.4%, respectively), and otologicals (62.2% versus 68.1%, respectively) (Table 15).

In the COMMODORE 1 study, the majority of patients in the crovalimab and eculizumab arms had at least 1 previous medical condition (75.6% versus 72.7%, respectively).¹⁷ The most frequent previous medical conditions were hypertension (15.6% in the crovalimab arm versus 27.3% in the eculizumab arm), cholelithiasis (8.9% versus 11.4%, respectively), and hypothyroidism (8.9% versus 6.8%, respectively).¹⁷ The proportion of patients who received at least 1 concomitant medication in the crovalimab arm and eculizumab arm was 93.3% and 86.4%, respectively. The most common concomitant medications were topical products for joint and muscular pain (73.3% in the crovalimab arm versus 72.7% in the eculizumab arm), ophthalmologicals (64.4% versus 61.4%, respectively) and antibacterials for systemic use (71.1% versus 50.0%, respectively) (Table 15).

At baseline, medications of special interest included immunosuppressive therapies, systemic corticosteroids, and antithrombotic drugs. Overall, the use of these selected therapeutics was comparable between the randomized arm A (crovalimab) and arm B (eculizumab) in the COMMODORE 2 study while in the COMMODORE 1 study, the rate of some medications in arm A was higher than in arm B: systemic corticosteroid therapy (34.1% versus 36.2%, respectively, in the COMMODORE 2 trial and 11.1% versus 6.8%, respectively, in the COMMODORE 1 trial), immunosuppressants (17.0% versus 18.8%, respectively, in the COMMODORE 2 trial and 42.2% versus 27.3%, respectively, the COMMODORE 1 trial), ██████████ ████ █████████ █ █████ █████ ███ ██████, and anticoagulants (25.9% versus 24.6%, respectively, in the COMMODORE 2 trial and 26.7% versus 22.7%, respectively, in the COMMODORE 1 trial).^15,17

In the COMMODORE 2 study, arm C (pediatric), systemic antibiotic therapeutics, immunosuppressants, and erythrocyte-stimulating drugs were administered to 1 pediatric patient each before enrolment. One patient received systemic antibacterial therapeutics on day 1.¹⁵

In the COMMODORE 1 study, arm C, at baseline, the proportion of patients receiving immunosuppressants or antithrombotic drugs was 63.2% and 15.8%, respectively. No patients received corticosteroids for systemic use at baseline.¹⁷

The majority of patients was immunized with Neisseria meningitidis serotypes A, C, W, and Y, with rates in the COMMODORE 2 study for arm A (crovalimab, 96.3%), arm B (eculizumab, 98.6%), and arm C (pediatric, 100%), and in the COMMODORE 1 study for arm A (crovalimab, 91.1%), arm B (eculizumab, 95.5%), and arm C (nonrandomized, 92.1%).^15,17

Table 15: Summary of Previous or Concomitant Treatment From the COMMODORE 2 Study (Arm A and Arm B) and the COMMODORE 1 Study (All Arms), Full Reporting Period

Note: Data presented in Table 15 were based on analyses at the clinical cut-off date of November 16, 2022.

Sources: COMMODORE 2 Primary Clinical Study Report¹⁵ and COMMODORE 1 Primary Clinical Study Report.¹⁷

Efficacy

A summary of the key efficacy results for the randomized comparison of crovalimab (arm A) versus eculizumab (arm B) in the primary treatment periods of the COMMODORE 2 and COMMODORE 1 studies is shown in Table 16 and Figure 12 (Appendix 1). The efficacy analyses included all randomized patients in arm A and arm B who were recruited at least 24 weeks before the CCOD of November 16, 2022.^15,17 Exploratory efficacy data from arm B patients switching from eculizumab to crovalimab beyond week 25 of the trials are summarized under each efficacy outcome for the time period under crovalimab treatment, following the switch. Depending on the availability and relevance of data, the results of the extension period are reported for arm B switch patients (68 patients in the COMMODORE 2 study and 35 patients in the COMMODORE 1 study) and/or arm B switch patients who switched to crovalimab at least 24 weeks before the data cut-off (43 patients in the COMMODORE 2 study and 28 patients in the COMMODORE 1 study).^15,17

Hemolysis Outcomes

Hemolysis Control

COMMODORE 2 study: Crovalimab demonstrated noninferiority compared with eculizumab treatment for hemolysis control as measured by a central LDH value of 1.5 or less multiplied by ULN from week 5 through week 25. The OR for hemolysis control (crovalimab versus eculizumab) was 1.02 (95% CI, 0.57 to 1.82). The lower limit of the 95% CI for the OR of 0.57 was greater than the predefined NIM of 0.2. The mean proportion of patients with hemolysis control as measured by a central LDH value of 1.5 or less multiplied by ULN from week 5 through week 25 was similar between the crovalimab arm at 79.3% (95% CI, 72.9% to 84.5%) and the eculizumab arm at 79.0% (95% CI, 69.7% to 86.0%; P = 0.95).

The proportion of patients who achieved hemolysis control increased from 0% in both treatment arms at baseline to 81.0% of patients (95% CI, 73.0% to 87.4%) in the crovalimab arm and 83.8% of patients (95% CI, 72.9% to 91.6%) in the eculizumab arm at week 5. These proportions remained between 75.2% and 83.8% in the crovalimab arm and between 73.8% and 85.1% in the eculizumab arm through week 25 (Figure 3). Overall, the results of the different sensitivity analyses using intention-to-treat and per-protocol populations and different statistical models were consistent with the primary analysis of hemolysis control. Specifically, a total of 7.5% missing central LDH values was reported from week 5 through week 25 (9.5% in the crovalimab arm versus 3.7% in the eculizumab arm). The results from the respective imputation-based sensitivity analyses (MCMC algorithm, predictive mean matching algorithm, and tipping point analysis) were consistent with those from the primary analysis.^15,23

Point estimates in the crovalimab arm were generally consistent across subgroups for hemolysis control. █████████ █████████ ██████ ██ ██████████ ██████ ██████████ ███ ████ ██ ███ █████ █████████ ███ █████████ ███████ ██ ███████ █████ █████ █████████ ███████████ █████ ███ ███ ████ ██ █████ ███ ██ █████ █████ ██████ █████ ███ ███ ████ ██ ██████.¹⁵ A different treatment effect by sex was not observed for the coprimary end point of transfusion avoidance. The sponsor reported that differences based on sex were not expected.

In the COMMODORE 2 trial, following the 24-week randomized, primary treatment period, of the 68 eculizumab-treated patients who switched to crovalimab (arm B switch patients), 50 patients (proportion of patients = 73.5%; 95% CI, 61.4% to 83.5%) had hemolysis control at the time of the switch (switch baseline). From switch baseline to 25 weeks later (switch week 25), the proportion of patients with hemolysis control ranged from 77.3% to 89.7% at each visit (Appendix 1, Figure 13).^15,23

Figure 3: COMMODORE 2 Study — Proportion of Patients With Hemolysis Control (Central LDH Value ≤ 1.5 × ULN) Through Week 25 by Visit, With 95% CI (Primary Analysis Population)

CI = confidence interval; Crova = crovalimab; Ecu = eculizumab; LDH = lactate dehydrogenase; pt = patient; ULN = upper limit of normal.

Note: Data shown in Figure 3 were based on analyses at the clinical cut-off date of November 16, 2022. For each group, CIs are displayed only for visits with at least 10 patients. The baseline LDH value was defined as the mean of all central LDH values, collected within 28 days before the first on-study drug administration including the predose value from day 1.

Source: COMMODORE 2 Primary Clinical Study Report.¹⁵

COMMODORE 1 study: The mean proportion of patients who achieved hemolysis control (a central LDH value ≤ 1.5 × ULN) from baseline through week 25 was 92.9% (95% CI, 86.6% to 96.4%) for the crovalimab arm versus 93.7% (95% CI, 87.3% to 97.0%) for the eculizumab arm (Table 16). The mean proportion of patients with a central LDH value of 1.5 or less multiplied by ULN at baseline was 93.2% in the crovalimab arm versus 95.2% in the eculizumab arm. These proportions ranged from 76.9% to 100.0% of patients in the crovalimab arm and 86.1% to 97.2% of patients in the eculizumab arm up to week 25.¹⁷

Of the 35 arm B patients who switched to crovalimab (arm B switch patients), 31 patients (proportion of patients = 88.6%; 95% CI, 73.3% to 96.8%) had hemolysis control at switch baseline. The proportion of patients with hemolysis control ranged between 89.7% and 100.0% across all following visits up to switch week 25.¹⁷ In the arm B switch patients who switched to crovalimab for at least 24 weeks (n = 28), the mean proportion of patients who achieved hemolysis control by the CCOD was 95.6% (95% CI, 87.3% to 98.6%).¹⁷

Breakthrough Hemolysis

COMMODORE 2 study: The proportion of patients with a BTH event from baseline through week 25 was 10.4% (95% CI, 6.0% to 17.2%) in the crovalimab arm compared with 14.5% (95% CI, 7.5% to 25.5%) in the eculizumab arm. The weighted difference in the proportions of patients with BTH (crovalimab versus eculizumab) was −3.9% (95% CI, −14.8% to 5.3%). The upper limit of the 95% CI for the difference in the proportions was 5.3%, which is lower than the predefined NIM of 20% (Table 16).

Of the 43 arm B switch patients, a BTH event was attributed to 7 patients (proportion of patients = 16.3%; 95% CI, 7.3% to 31.3%) between switch baseline and switch week 25. Of these, 4 patients had recorded a BTH event while the other 3 patients were considered to have had a BTH event as a conservative analysis approach.¹⁵

COMMODORE 1 study: The proportion of patients with BTH from baseline through week 25 was 10.3% (95% CI, 3.3% to 25.2%) for the crovalimab arm versus 13.5% (95% CI, 5.1% to 29.6%) for the eculizumab arm (Table 16). Two patients in the eculizumab arm without a BTH event discontinued treatment before week 25 and were therefore conservatively assumed as having had a BTH event.

Of the 28 arm B switch patients who switched to crovalimab for at least 24 weeks, a BTH event was attributed to 5 (17.9%) patients between switch baseline and switch week 25. Of these, 3 patients were considered to have had a BTH event as a conservative analysis approach.¹⁷

Mean Percentage Change in LDH Levels

COMMODORE 2 study: The mean percentage reduction in LDH levels (central assessment) from baseline to week 25 was −73.6% (95% CI, −79.0% to −68.2%) in the crovalimab arm compared with −64.1% (95% CI, −71.4% to −56.8%) in the eculizumab arm (Table 16).

Results of this outcome were not reported for the arm B switch patients in the extension period.

COMMODORE 1 study: The mean percentage change in the central LDH value from baseline to the average over week 21, week 23, and week 25 was 16.6% (95% CI, ███ ██ ████) for the crovalimab arm versus 4.5% (95% CI, ████ ██ ████) for the eculizumab arm (Table 16). While the percentage change (increase) in the central LDH value from baseline was higher in the crovalimab arm compared with the eculizumab arm, the mean central LDH value remained at 1.5 or less multiplied by ULN for both the crovalimab and eculizumab arms throughout the primary treatment period (Appendix 1, Figure 14) and up to the CCOD. The mean proportion of patients with a central LDH value of 1.5 or less multiplied by ULN by visit through week 25 was also similar between the 2 arms (refer to Table 16).^17,23

Results of this outcome were not reported for the arm B switch patients in the extension period.

Hemoglobin Outcomes

Stabilized Hemoglobin

COMMODORE 2 study: The proportion of patients reaching hemoglobin stabilization (the avoidance of a ≥ 2 g/dL decrease in hemoglobin level from baseline, in the absence of transfusion) from baseline through week 25 was 63.4% (95% CI, 54.6% to 71.5%) in the crovalimab arm compared with 60.9% (95% CI, 48.4% to 72.2%) in the eculizumab arm. The weighted difference in the proportion of patients with hemoglobin stabilization (crovalimab versus eculizumab) was 2.2% (95% CI, −11.4% to 16.3%) and the lower limit of the 95% CI of −11.4% was higher than the predefined NIM of −20% (Table 16).

In arm B switch patients, 27 of 43 patients (proportion of patients = 62.8%; 95% CI, 46.7% to 76.6%) achieved hemoglobin stabilization from switch baseline to switch week 25. Of these, 2 patients who were hemoglobin-stabilized while on study discontinued treatment before week 25 and were therefore conservatively assumed as not having stabilized hemoglobin.¹⁵

COMMODORE 1 study: The proportion of patients with stabilized hemoglobin from baseline through week 25 was 59.0% (95% CI, 42.2% to 74.0%) in the crovalimab arm versus 70.3% (95% CI, 52.8% to 83.6%) in the eculizumab arm (Table 16).

A total of 18 of the 28 (64.3%) arm B switch patients in the 24-week crovalimab efficacy population achieved hemoglobin stabilization from switch baseline to switch week 25.

Patients Who Reached and Maintained a Minimum Hemoglobin Level (Hemoglobin ≥ 10 g/dL Without Subsequent Decrease < 9 g/dL)

In the COMMODORE 2 study, the proportion of patients who reached and maintained a hemoglobin level of at least 10 g/dL (without a subsequent decrease below 9 g/dL, in the absence of transfusion) from baseline through week 25 was █████ ████ ███ ████ ██ █████ in the crovalimab arm and █████ ████ ███ ████ ██ █████ in the eculizumab arm.

In the COMMODORE 1 study, the proportion of patients who reached and maintained a hemoglobin level of at least 10 g/dL (without a subsequent decrease below 9 g/dL, in the absence of transfusion) from baseline through week 25 was █████ ████ ███ ████ ██ █████ in the crovalimab arm versus █████ ████ ███ ████ ██ █████ in the eculizumab arm.

In both the COMMODORE 2 and COMMODORE 1 trials, results of this outcome were not reported for the arm B switch patients in the extension period.

Transfusion Outcomes

Transfusion Avoidance

COMMODORE 2 study: Crovalimab demonstrated noninferiority compared with eculizumab for transfusion avoidance from baseline through week 25. The weighted difference in the proportion of patients with transfusion avoidance (crovalimab versus eculizumab) was −2.8% (95% CI, −15.7% to 11.1%; P = ███) with a lower limit of the 95% CI of −15.7%, which was higher than the predefined NIM of −20% (Table 16). In the crovalimab arm, 65.7% (95% CI, 56.9% to 73.5%) of patients were transfusion-free from baseline through week 25 compared with 68.1% (95% CI, 55.7% to 78.5%) of patients in the eculizumab arm. One patient in the crovalimab arm discontinued treatment before week 25 without a transfusion and was conservatively assumed to have had a transfusion. Overall, the results of the sensitivity analyses using different analysis population definitions were consistent with the results from the primary analysis.^15,23

Of the 43 arm B switch patients, 33 patients (proportion of patients = 76.7%; 95% CI, 61.0% to 87.7%) achieved transfusion avoidance in the period after switching to crovalimab through switch week 25. Two arm B switch patients discontinued treatment before switch week 25 without a transfusion and were conservatively assumed to have had a transfusion.^15,23

Point estimates in the crovalimab arm were generally consistent across subgroups for transfusion avoidance.¹⁵

COMMODORE 1 study: The proportion of patients who achieved transfusion avoidance from baseline through week 25 was 79.5% (95% CI, 63.1% to 90.1%) for the crovalimab arm versus 78.4% (95% CI, 61.3% to 89.6%) for the eculizumab arm (Table 16). One patient in the eculizumab arm discontinued treatment before week 25 without a transfusion and was conservatively assumed as having had a transfusion.

Of the 28 arm B switch patients who switched to crovalimab for at least 24 weeks, 23 (82.1%) patients achieved transfusion avoidance after switching to crovalimab through switch week 25.¹⁷

Red Blood Cell Transfusions

COMMODORE 2 study: Among the primary analysis population, the mean number of units of pRBC transfused from baseline through week 25 in the primary analysis population was similar between the crovalimab arm at 2.3 units (95% CI, 1.3 units to 3.4 units) and the eculizumab arm at 2.2 units (95% CI, 1.0 unit to 3.4 units) (Table 16). During the primary treatment period, 45 (33.6%) patients in the crovalimab arm and 22 (31.9%) patients in the eculizumab arm had at least 1 transfusion. Among patients who received at least 1 transfusion, the mean number of units of pRBC transfused was similar with 6.9 units (95% CI, 4.3 units to 9.6 units) in the crovalimab arm and 6.9 units (95% CI, 4.1 units to 9.8 units) in eculizumab arm.

In the 43 arm B switch patients who switched to crovalimab for at least 24 weeks, 8 (18.6%) patients had at least 1 transfusion from switch baseline up to switch week 25. The mean number of units of pRBC transfused among these 43 patients was 1.9 units (95% CI, 0.4 unit to 3.4 units) from switch baseline through switch week 25.^15,23 Among the arm B switch patients with at least 1 transfusion (n = 8), the mean number transfused was 10.3 units (95% CI, 4.9 units to 15.6 units).¹⁵

COMMODORE 1 study: Among the efficacy-evaluable population, 8 (20.5%) patients in the crovalimab arm and 7 (18.9%) patients in the eculizumab arm had at least 1 transfusion, and the mean number of pRBC units transfused from baseline to week 25 in all randomized patients was 0.97 unit (95% CI, 0.24 unit to 1.70 units) in the crovalimab arm and 1.89 units (95% CI, 0.53 unit to 3.25 units) in the eculizumab arm (Table 16). The mean units transfused were not reported among patients who had at least 1 transfusion.

The mean number of units of pRBC transfused among the 28 arm B switch patients who switched to crovalimab for at least 24 weeks was 1.0 (SD = 2.96) unit(s).¹⁷

Patient-Reported Outcomes

Functional Assessment of Chronic Illness Therapy–Fatigue

In both the COMMODORE 2 and COMMODORE 1 studies, FACIT-F was assessed in adult patients only (≥ 18 years).

Table 16: Summary of Key Efficacy Results From the COMMODORE 2 Study (Primary Analysis Population) and COMMODORE 1 Study (24-Week Efficacy Population)

BTH = breakthrough hemolysis; CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FACIT-F = Functional Assessment of Chronic Illness Therapy–Fatigue; GEE = generalized estimating equation; GHS = global health status; Hb = hemoglobin; LDH = lactate dehydrogenase; MMRM = mixed model of repeated measures; NIM = noninferiority margin; NR = not reported; PNH = paroxysmal nocturnal hemoglobinuria; pRBC = packed red blood cell; QoL = quality of life; RCT = randomized controlled trial; SD = standard deviation; SE = standard error; ULN = upper limit of normal.

Note: Data presented in Table 16 were based on analyses at the clinical cut-off date of November 16, 2022. In the COMMODORE 2 study, the outcomes of patients who reached and maintained a minimum Hb level, a mean percentage change in LDH levels, mean units of pRBC transfused, and an EORTC QLQ-C30 GHS and QoL score were not adjusted for multiplicity. In the COMMODORE 1 study, no adjustment for multiplicity was conducted for all outcomes. Definitions of the key efficacy outcomes are as follows. BTH was defined as at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, shortness of breath [dyspnea], anemia [Hb < 10 g/dL], a major adverse vascular event [including thrombosis], dysphagia, or erectile dysfunction) in the presence of elevated LDH of 2 or more multiplied by ULN after the prior reduction of LDH to 1.5 or less multiplied by ULN on treatment. Stabilized Hb was defined as avoidance of a 2 g/dL or greater decrease in Hb level from baseline, in the absence of transfusion. Patients who reached and maintained a minimum Hb level were defined as patients who reached an Hb level of at least 10 g/dL, without a subsequent decrease below 9 g/dL, in the absence of transfusion. Transfusion avoidance was defined as patients who were pRBC transfusion–free and did not require a transfusion per protocol-specified guidelines. The FACIT-F tool consists of 13 items that assess fatigue using a 7-day recall period. Items are scored on a response scale that ranges from 0 (“not at all”) to 4 (“very much so”); relevant items are reverse-scored and all items are summed to create a total score, with a higher score indicative of better functioning (i.e., less fatigue). The total score ranges from 0 to 52. MMRMs were used in analysis. The EORTC QLQ-C30 GHS and QoL scale was assessed in adult patients only; the scale ranges from 0 to 100, with higher scores indicating better GHS and QoL.

^aEstimates were based on a GEE model.

^bAn odds ratio of greater than 1 favours crovalimab.

^cThe P value was based on the Mantel-Haenszel test.

^dThe predefined statistical testing hierarchy was broken before superiority testing could be conducted. Therefore, the P value reported is descriptive only.

^eFour patients in the crovalimab arm and 1 patient in the eculizumab arm of the COMMODORE 2 study, as well as 2 patients in the eculizumab arm of the COMMODORE 1 study, discontinued their respective studies before week 25 and were considered to have experienced a BTH event as a conservative analysis approach.

^fOne patient in the crovalimab arm of the COMMODORE 2 study discontinued the study before week 25 with Hb stabilization and was conservatively assumed to have not had Hb stabilization.

^gOne patient in the crovalimab arm of the COMMODORE 2 study and 1 patient in the eculizumab arm of the COMMODORE 1 study discontinued their respective studies before week 25 without a transfusion and were conservatively assumed to have had a transfusion.

^hFACIT-F was assessed in adult patients only (crovalimab arm = 134 adult patients; eculizumab arm = 67 adult patients). The total FACIT-F score ranges from 0 to 52 with higher scores indicating lower fatigue severity. The threshold for a clinically meaningful change is 5 points or more.

ⁱNoninferiority testing of FACIT-F was planned to occur only after successful superiority testing of all the other coprimary and secondary efficacy end points. Given the outcome of this superiority testing, the comparative results of FACIT-F are descriptive only. The summary of MMRM FACIT-F scores are presented in Table 16.

Sources: COMMODORE 2 Primary Clinical Study Report,¹⁵ COMMODORE 1 Primary Clinical Study Report,¹⁷ and sponsor’s submission.¹⁶

COMMODORE 2 study: Due to the break in the statistical testing hierarchy, noninferiority testing for FACIT-F was not performed and the results were considered descriptive only. FACIT-F data were evaluable in 95.5% of adult patients in the crovalimab arm and 95.7% of adult patients in the eculizumab arm at each visit from baseline through week 25. The mean FACIT-F scores at baseline were below normative values and similar in the crovalimab arm at 36.0 points (95% CI, 34.29 points to 37.76 points) and in the eculizumab arm at 35.1 points (95% CI, 32.28 points to 37.90 points) (Table 16 as well as Appendix 1, Figure 15). Improvement in fatigue was observed by week 2, with the mean score increasing to 40.9 points (95% CI, 39.57 points to 42.28 points) in the crovalimab arm compared with 38.9 points (95% CI, 36.35 points to 41.47 points) in the eculizumab arm. Further improvement in levels of fatigue were reported up to week 25, with the mean score increasing to 44.3 points (95% CI, 43.15 points to 45.52 points) in the crovalimab arm compared with 41.4 points (95% CI, 39.29 points to 43.47 points) in the eculizumab arm. The adjusted mean change from baseline at week 25 in the FACIT-F score was 7.8 points (95% CI, 6.49 points to 9.09 points) in the crovalimab arm compared with 5.2 points (95% CI, 3.42 points to 6.89 points) in the eculizumab arm.

For the proportion of patients with a 5-point or greater improvement from baseline in the FACIT-F score, improvement of fatigue was observed by week 2, with 41.7% of patients (95% CI, 33.25% to 50.57%) in the crovalimab arm and 40.3% of patients (95% CI, 28.72% to 53.00%) in the eculizumab arm exceeding the threshold for clinically meaningful change (≥ 5 points) (Appendix 1, Table 34). By week 25, 58.6% of patients in the crovalimab arm and 54.5% of patients in the eculizumab arm had an improvement in fatigue severity of at least 5 points.¹⁵

In 36 of the 43 arm B switch patients who switched to crovalimab for at least 24 weeks and with available data for this outcome, the adjusted mean change from switch baseline to switch week 25 in the FACIT-F score was 0.18 points (95% CI, −2.72 points to 3.09 points).¹⁵

COMMODORE 1 study: FACIT-F data were evaluable in 86.4% of adult patients in the crovalimab arm (n = 38) and 76.2% of adult patients in the eculizumab arm (n = 32) at each visit from baseline through week 25. The adjusted mean change in FACIT-F scores from baseline to week 25 was 1.1 points (95% CI, −1.47 points to 3.65 points) for the crovalimab arm and −2.6 points (95% CI, −5.35 points to 0.12 point) for the eculizumab arm (Table 16 as well as Appendix 1, Figure 16).

The proportion of patients with a 5-point or greater improvement from baseline in the FACIT-F score was not reported for the COMMODORE 1 study.^17,23

In 26 of the 35 arm B switch patients who had completed the assessment at both visits, the adjusted mean change from switch baseline to switch week 25 in the FACIT-F score was 2.50 (95% CI, 0.03 to 4.98).¹⁷

EORTC QLQ-C30 and EORTC IL-40 Populations

EORTC QLQ-C30 and EORTC IL-40 results were evaluable in:

• the COMMODORE 2 study — 95.5% of adult patients in arm A (crovalimab) (n = 128) and 95.7% of adult patients in arm B (eculizumab) (n = 66) at each visit from baseline through week 25

• the COMMODORE 1 study — 86.4% of adult patients in arm A (crovalimab) (n = 38) and 76.2% of adult patients in arm B (eculizumab) (n = 32) at each visit from baseline through week 25.

EORTC QLQ-C30 GHS and QoL

COMMODORE 2 study: The mean score at baseline was comparable between the crovalimab arm at 62.7 points (95% CI, 59.34 points to 66.03 points) and the eculizumab arm at 62.4 points (95% CI, 57.59 points to 67.28 points), indicating moderate levels of GHS and QoL. The mean change from baseline to week 25 was 13.4 points (95% CI, 10.11 points to 16.72 points) for the crovalimab arm and 9.9 points (95% CI, 4.82 points to 14.88 points) for the eculizumab arm. Mean values at week 25 were similar to normative population values.^15,65,66

From switch baseline to switch week 25, the GHS and QoL scores remained relatively stable in arm B switch patients of the crovalimab efficacy population.¹⁵

COMMODORE 1 study: The mean change in EORTC GHS and QoL score from baseline to week 25 was 5.7 (95% CI, −2.36 to 13.76) in the crovalimab arm and −1.0 (95% CI, −6.94 to 4.86) in the eculizumab arm.¹⁷

Up to switch week 25, the GHS and QoL scores remained relatively stable in the 26 arm B switch patients of the crovalimab efficacy population.¹⁷

EORTC Item Library 40

COMMODORE 2 study: Overall, there was an improvement in dyspnea, dysphagia, chest pain, abdominal pain, and erectile dysfunction scores after starting treatment with either crovalimab or eculizumab, up to week 25 (Appendix 1, Table 35). In both treatment arms, the EORTC IL-40 headache score was numerically larger at week 2 compared to baseline, with a change from baseline of 4.3 (SD = 29.2; 95% CI, −0.74 to 9.33) in the crovalimab arm and 9.5 (SD = 25.8; 95% CI, 3.15 to 15.75) in the eculizumab arm. However, the scores decreased from week 5 onward, and the change from baseline at week 25 in the crovalimab arm was −6.8 (SD = 25.2; 95% CI, −11.18 to −2.36) and in the eculizumab arm was −4.6 (SD = 18.4; 95% CI, −9.06 to −0.03).¹⁵

From switch baseline to switch week 25, all symptoms assessed using the EORTC IL-40 tool (dyspnea, dysphagia, chest pain, headaches, abdominal pain, and erectile dysfunction) remained relatively stable in the arm B switch patients.¹⁵

COMMODORE 1 study: Overall, the point estimates suggested there was little to no difference in dyspnea, dysphagia, chest pain, headaches and abdominal pain scores from baseline to week 25 in both the crovalimab and eculizumab arms (Appendix 1, Table 36). Erectile dysfunction scores were numerically larger (worsening) from baseline to week 25 in the crovalimab arm (change in mean [SD] = 6.7 [38.2]) and the eculizumab arm (change in mean [SD] = 7.1 [32.5]).¹⁷

In arm C, up to week 25, the outcome data suggested little to no difference in point estimates of dysphagia, chest pain and abdominal pain assessed by EORTC IL-40 in the patients from the prior ravulizumab, prior high-dose eculizumab, and C5 polymorphism arm C cohorts from baseline to week 25. The study results showed a small numeric increase (worsening) in dyspnea scores and a small numeric decrease (improvement) in headache scores compared with baseline in the cohort of patients who switched from ravulizumab. The sample size of male patients (n = 8 in the prior ravulizumab cohort, n = 3 in the prior high-dose eculizumab cohort, and n = 2 in the C5 polymorphism cohort) was too low to draw firm conclusions from the summary data for erectile dysfunction.^17,23

From switch baseline to switch week 25, all symptoms assessed using the EORTC IL-40 tool (dyspnea, dysphagia, chest pain, headaches, abdominal pain, and erectile dysfunction) remained relatively stable in arm B switch patients.¹⁷

Proportion of Patients With a Preference for Crovalimab per PPQ

COMMODORE 2 study: At switch week 17, 84% of the arm B switch patients (32 of 38 patients) in the crovalimab efficacy population who switched from eculizumab treatment to crovalimab treatment preferred treatment with crovalimab. The remaining patients preferred eculizumab (10.5%) or had no preference (5.3%) (Table 17). The top 3 reasons for crovalimab preference were “(t)he way treatment was given was easier,” “(f)ewer hospital visits associated with treatment,” and “(t)ime to administer treatment was shorter.”¹⁵

COMMODORE 1 study: At week 17 of treatment, 85% of patients in the crovalimab arm (33 of 39 patients) preferred treatment with crovalimab, 2.6% of patients preferred eculizumab, and 12.8% of patients had no preference (Table 17). The top 3 reasons given for patient preference were “(t)he way treatment was given was easier,” “(t)ime to administer treatment was shorter,” and “(f)ewer hospital visits associated with treatment.”¹⁷

At switch week 17, most of the arm B switch patients (27 of 28 [96.4%] patients) in the crovalimab efficacy population who switched from eculizumab treatment to crovalimab treatment preferred treatment with crovalimab (Table 17). The remaining patient had no preference between eculizumab and crovalimab treatments. The top 3 reasons for preference for crovalimab were “(f)ewer hospital visits associated with treatment,” “(b)etter quality of life,” and “(t)he way treatment was given was easier.”¹⁷

Table 17: Proportion of Patients With a Preference for Crovalimab or Eculizumab (Arm B, Period After Switch to Crovalimab in the COMMODORE 2 Study, and Arm A and Arm B, Primary Efficacy Period in the COMMODORE 1 Study)

Note: Data presented in Table 17 were based on analyses at the clinical cut-off date of November 16, 2022.

^aPatients who were initially randomized to eculizumab had the opportunity to switch to crovalimab in the extension period after completing 24 weeks of eculizumab treatment.

^bOnly participants with available data (having completed the questionnaire) were included in the calculations of percentages.

^cPatients were under treatment with eculizumab before enrolment and were randomized to arm A, crovalimab.

Sources: COMMODORE 2 Primary Clinical Study Report,¹⁵ COMMODORE 1 Primary Clinical Study Report,¹⁷ and sponsor’s submission.²³

At week 17, 57% of the patients in arm C who were pretreated with eculizumab (the higher-than-approved dose eculizumab or C5 polymorphism patients) preferred treatment with crovalimab (8 of 14 patients). Smaller proportions of patients had no preference (4 of 14 [28.6%] patients) or preferred eculizumab (2 of 14 [14.3%] patients). At week 17, 60% of the patients in arm C pretreated with ravulizumab preferred treatment with crovalimab (9 of 15 patients). Smaller proportions had no preference (2 of 15 [3.3%] patients) or preferred ravulizumab (4 of 15 [26.7%] patients).¹⁷

Summary of Efficacy in Arm C of COMMODORE 2 and COMMODORE 1 Studies

COMMODORE 2 Study, Arm C (Nonrandomized, Descriptive, and Pediatric [Patients < 18 Years])

Hemolysis control: In the 6 pediatric patients, the normalized LDH value (central assessment) ranged between 3.8 and 23.4 multiplied by ULN at baseline. All 6 patients reached a central LDH value of 1.5 or less multiplied by ULN between week 2 and week 4 of treatment with crovalimab, and this was sustained across the first 24 weeks of treatment for all except 1 patient (Appendix 1, Figure 17). In this 1 patient, a central LDH value of 1.5 or less multiplied by ULN was reached by week 4; however, starting at week 7, the LDH value increased to 1.6 multiplied by ULN and thereafter ranged between 1.41 and 1.81 multiplied by ULN through to week 25.¹⁵

Transfusion avoidance: Four of the 6 pediatric patients achieved transfusion avoidance from baseline through to week 25. During the primary treatment period, 2 patients had at least 1 transfusion postbaseline. One patient with a documented history of aplastic anemia had a pRBC transfusion at a single time point on day 2. A second patient with a history of aplastic anemia and a history of transfusion in the 12 months before study start (with 40.5 pRBC units transfused in the prior 12 months) had multiple transfusions on study starting at day 24.¹⁵

BTH: No BTH events were observed in the 6 pediatric patients from baseline to week 25.¹⁵

Stabilized hemoglobin: Four of 6 pediatric patients achieved hemoglobin stabilization from baseline to week 25. The 2 patients who did not achieve hemoglobin stabilization were the same 2 patients who received a transfusion during the primary treatment period.¹⁵

Mean change in PedsQL MFS and physical functioning scale of the PedsQL Generic Core Scales: PedsQL Generic Core Scales and PedsQL MFS were evaluable in all 6 pediatric patients at each visit from baseline through week 25.

Overall fatigue scores from the PedsQL MFS tool were variable across time points for all 6 pediatric patients. Most patients experienced an overall improvement in fatigue over the course of treatment. One patient had a similar score at their last time point (week 49) compared to baseline.¹⁵

Physical functioning scores were variable across time points for all 6 pediatric patients. Most patients experienced an overall improvement in physical functioning. One patient had a similar score at their last time point (week 29) compared to baseline, and 1 patient experienced a decline in physical functioning.¹⁵

COMMODORE 1 Study, Arm C (Nonrandomized Cohorts)

Analysis of exploratory efficacy end points included only patients who had had the opportunity to complete 24 weeks of treatment (i.e., were enrolled more than 24 weeks before CCOD). This corresponded to 19 patients in the prior ravulizumab cohort, 9 patients in the prior high-dose eculizumab cohort, and 6 patients in the C5 polymorphism cohort.¹⁷

The summary of exploratory efficacy outcomes from the COMMODORE 1 study are presented in Table 18.

Table 18: Summary of Exploratory Efficacy Outcomes From the COMMODORE 1 Study, Arm C (Nonrandomized Arm)

BTH = breakthrough hemolysis; C5 = complement component 5; CI = confidence interval; LDH = lactate dehydrogenase; pRBC = packed red blood cell; SNP = single nucleotide polymorphism; ULN = upper limit of normal.

^aTwo patients in the prior ravulizumab cohort and 1 patient in the high-dose eculizumab cohort did not have a transfusion on study but as a conservative analysis approach, they were regarded as having had a transfusion given that they discontinued the study treatment before week 25.

^bThree patients in the prior ravulizumab cohort and 1 patient in the prior high-dose eculizumab cohort without a BTH discontinued treatment before week 25 and were therefore conservatively considered to have had a BTH event.

^cTwo patients in the prior ravulizumab cohort and 1 patient in the prior high-dose eculizumab cohort discontinued crovalimab treatment before week 25 and as a conservative analysis approach were therefore regarded as not having had stabilized hemoglobin.

Source: COMMODORE 1 Primary Clinical Study Report.¹⁷

FACIT-F: Scores in patients from arm C cohorts suggested that there was little to no difference postbaseline up to week 25.¹⁷

EORTC QLQ-C30: Physical functioning, role functioning, and GHS and QoL scores suggested that there was little to no difference in patients from all arm C cohorts postbaseline up to week 25.¹⁷

EORTC IL-40: From baseline to week 25, dysphagia, chest pain, and abdominal pain scores suggested that there was little to no difference in in the patients from all arm C cohorts. The study results showed a small numeric increase (worsening) in dyspnea scores and a small numeric decrease (improvement) in headache scores as compared to baseline in the cohort of patients who switched from ravulizumab. The sample size of male patients for the assessment of erectile dysfunction was very small (n = 8 in the prior ravulizumab cohort, n = 3 in the prior high-dose eculizumab cohort, and n = 2 in the C5 polymorphism cohort).¹⁷

PPQ: At week 17, most patients in arm C who had been pretreated with eculizumab (high-dose eculizumab or C5 polymorphism patients) preferred treatment with crovalimab (8 of 14 [57.1%] patients). Smaller proportions of patients had no preference (4 of 14 [28.6%] patients) or preferred eculizumab (2 of 14 [14.3%] patients). At week 17, a similar proportion of patients in arm C pretreated with ravulizumab preferred treatment with crovalimab (9 of 15 [60.0%] patients). Smaller proportions of patients had no preference (2 of 15 [13.3%] patients) or preferred ravulizumab (4 of 15 [26.7%] patients).¹⁷

PedsQL for pediatric patients: One pediatric patient was enrolled approximately 2 weeks before the CCOD and therefore data for this patient were limited.¹⁷

Harms

A summary of harms reported in the randomized safety populations during the 24-week primary safety period in the COMMODORE 2 and COMMODORE 1 trials is provided in Table 19.

Adverse Events

In the COMMODORE 2 study, the proportion of patients with at least 1 AE was comparable between the crovalimab arm (77.8%) and the eculizumab arm (79.7%). Infusion-related reaction was the most commonly reported AE by preferred term in the crovalimab arm (15.6%). The most frequent AEs in the eculizumab arm were also infusion-related reaction, together with urinary tract infection, and hypokalemia (all 13.0%). Injection-related reactions occurred in the crovalimab arm (5.2%) but not in the eculizumab arm due to SC administration being unique to the crovalimab arm. The proportion of patients who experienced infusion-related reactions was comparable between the crovalimab arm (15.6%) and the eculizumab arm (13.0%). The proportion of patients who experienced at least 1 infection was lower in the crovalimab arm (23.7%) compared with the eculizumab arm (36.2%). No cases of infection with N. meningitidis, including meningococcal meningitis, were reported in either arm. Eight (5.9%) patients in the crovalimab arm versus no patients in the eculizumab arm reported hypersensitivity other than type III immune complex reactions.

In the COMMODORE 1 study, the proportion of patients with at least 1 AE in the crovalimab arm (77.3%) was higher than in the eculizumab arm (66.7%). This difference was driven by type III immune complex events, injection-related reactions, and infusion-related reactions. The most frequent AE in the randomized safety population was COVID-19, with 13.6% of patients in the crovalimab arm and 16.7% of patients in the eculizumab arm experiencing it. Infusion-related reactions occurred only in the crovalimab arm (13.6%) but not in the eculizumab arm as these reactions were less likely to occur in the eculizumab arm due to patients starting the study stabilized on eculizumab treatment. A similar proportion of patients experienced infections in the crovalimab arm (40.9%) and the eculizumab arm (35.7%), and there were no cases of meningococcal meningitis in either arm. Four (9.1%) patients in the crovalimab arm versus no patients in the eculizumab arm reported hypersensitivity other than type III immune complex reactions.

Serious Adverse Events

In the COMMODORE 2 study, a similar proportion of patients reported at least 1 SAE in the crovalimab arm (10.4%) and the eculizumab arm (13.0%). In total, 3.0% of patients in the crovalimab arm and 1.4% of patients in the eculizumab arm experienced SAEs considered by the investigators to be related to study treatment.¹⁵

In the COMMODORE 1 study, a higher proportion of patients experienced SAEs in the crovalimab arm (13.6%) compared with the eculizumab arm (2.4%). In the crovalimab arm, the most frequent SAEs by System Organ Class were infections and infestations (6.8%). There was no obvious pattern in the SAEs observed and none of the SAEs were considered by the investigators to be related to crovalimab or eculizumab.¹⁷

Withdrawals Due to Adverse Events

In the COMMODORE 2 study, 1 patient each from the crovalimab arm (0.7%) and the eculizumab arm (1.4%) experienced AEs leading to the discontinuation of treatment.

In the COMMODORE 1 study, no patients experienced AEs leading to the withdrawal from the crovalimab or eculizumab arms.

Mortality

In the COMMODORE 2 study, death was reported in 2 (1.5%) patients in the crovalimab arm and 1 (1.4%) patient in the eculizumab arm. No deaths were related to treatment per the study investigators’ assessment.¹⁵

In the COMMODORE 1 study, no deaths were reported during the primary safety period.

Notable Harms

In the COMMODORE 2 study, no AESIs of type III immune complex reactions related to DTDCs were reported in either arm during the primary treatment period because the patients were treatment-naive. No AESIs of abnormal liver function tests or suspected transmission of an infectious drug by the study drug were reported in either arm. No clinically significant changes from baseline in vital signs and ECGs were observed during treatment with crovalimab or eculizumab.¹⁵

In the COMMODORE 1 study, 7 (15.9%) patients in the crovalimab arm versus no patients in the eculizumab arm experienced an AESI of type III immune complex reactions, because type III immune complex reactions were expected only in patients who switched from eculizumab to crovalimab who were at risk of developing DTDC-associated type III immune complex reactions. There were no AESIs reported for abnormal liver function tests or suspected transmission of an infectious drug by the study drug. Overall, in both treatment arms, no clinically significant changes from baseline for vital signs and ECGs were observed.¹⁷

Percentage of Patients Experiencing MAVE From Baseline to Week 25

COMMODORE 2 Study

One patient on crovalimab and 1 patient on eculizumab had a MAVE during the study; both patients had a prior history of MAVEs. One patient in the crovalimab arm had a MAVE of a myocardial infarction on study day 1 and subsequently died on study day 2 as a consequence of this AE. Retrospective analysis of laboratory data indicated that this patient had a myocardial infarction before crovalimab dosing. One patient in the eculizumab arm had a MAVE of a transient ischemic attack on study day 49. This AE occurred 6 days after the last dose of study drug administration (on day 43). The patient received treatment for this AE but subsequently died on day 71. In the investigators’ opinion, neither of the AEs in the 2 arms was related to treatment with the study drug.¹⁵

No MAVEs were reported in arm B switch patients from switch baseline through switch week 25.¹⁵

COMMODORE 1 Study

No patients in the crovalimab arm and 3 patients in the eculizumab arm were regarded as having a MAVE during the primary treatment period from baseline to week 25. In the eculizumab arm, 1 of the 3 patients experienced a MAVE (a transient ischemic attack). This patient experienced this MAVE on study day 20. The patient received treatment and the event resolved on study day 21. This patient did not have a prior history of a MAVE before entering the study. In the investigators’ opinion, this AE was not related to treatment with eculizumab. The other 2 patients in the eculizumab arm were conservatively considered as having a MAVE due to treatment discontinuation before 24 weeks without an actual MAVE reported.¹⁷

Table 19: Summary of Harms Results From the COMMODORE 2 and COMMODORE 1 Studies (24-Week Primary Safety Period, Randomized Safety Population)

AE = adverse event; AESI = adverse event of special interest; CCOD = clinical cut-off date; CI = confidence interval; DTDC = drug-target-drug complex; MAVE = major adverse vascular event; NA = not applicable; NR = not reported; RD = relative difference; SAE = serious adverse event.

Note: Data presented in Table 19 were based on analyses at the CCOD of November 16, 2022. Only treatment-emergent AEs are displayed.

^aInjection-related reactions were part of the “hypersensitivity other than type III immune complex reactions” group preferred term classification approach.

^bDTDCs only occur in patients switching to or from other C5 inhibitors, which bind to different epitopes than crovalimab does, and therefore were not relevant or expected in this part of the COMMODORE 2 study. Accordingly, no AESIs of type III immune complex reactions related to DTDCs were reported during the primary safety period in the randomized safety population that was treatment-naive or in patients who continued in the crovalimab arm after the primary safety period up to the CCOD.

^cMAVE was an exploratory end point being assessed in the primary analysis population in the COMMODORE 2 study. One of the 135 patients in arm A (crovalimab) was excluded from the primary analysis population for not fulfilling the criteria of at least 1 postbaseline LDH level assessment.

Sources: COMMODORE 2 Primary Clinical Study Report,¹⁵ COMMODORE 1 Primary Clinical Study Report,¹⁷ and sponsor’s submission.¹⁶

Summary of Harms in Arm C of COMMODORE 2 and COMMODORE 1 Studies

In the COMMODORE 2 study, overall, 83.3% (5 of 6) of pediatric patients experienced at least 1 AE (all were grade 1 to grade 2 in intensity). No deaths were reported. No pediatric patients experienced SAEs. One (16.7%) pediatric patient experienced a treatment-related AE of fatigue (grade 1), which resolved without treatment and dose modification. No pediatric patients had an AE leading to the withdrawal of treatment or dose modification. No pediatric patients experienced selected AEs (injection site reactions, infusion-related reactions, infections, and hypersensitivity reactions other than type III immune complex reactions) or AESIs. No clinically significant changes from baseline for vital signs and ECGs were observed. No MAVEs were reported in the 6 pediatric patients from baseline through week 25.¹⁵

In the COMMODORE 1 study, overall, 18 (85.7%) patients in the prior ravulizumab cohort, 10 (100%) patients in the prior high-dose eculizumab cohort, and 5 (83.3%) patients in the C5 polymorphism cohort experienced at least 1 AE. The single patient enrolled in the pediatric cohort (enrolled approximately 2 weeks before CCOD) did not experience any AEs. No deaths were reported in arm C. There was 33.3% of patients in the prior ravulizumab cohort and 20.0% of patients in the prior high-dose eculizumab cohort who experienced at least 1 SAE. No patients in the C5 polymorphism cohort experienced SAEs. One (4.8%) patient in the prior ravulizumab cohort had a grade 3 AE of sepsis leading to the withdrawal of treatment. Selected AEs of injection site reactions, infusion-related reactions, and hypersensitivity other than type III immune complex reactions were experienced in a small number of patients across cohorts. Infections were experienced in 42.9% (9) of patients, 50.0% (5) of patients, and 50.0% (3) of patients in the prior ravulizumab, prior high-dose eculizumab, and C5 polymorphism cohorts, respectively. There were no cases of meningococcal meningitis across all arm C cohorts. Overall, 23.8% of patients in the prior ravulizumab cohort and 20.0% of patients in the prior high-dose eculizumab cohort experienced at least 1 AESI of type III immune complex event. There were no AESIs reported for abnormal liver function tests or suspected transmission of an infectious drug by the study drug. Overall, no clinically significant changes from baseline for vital signs and ECGs were observed across all arm C cohorts. The number of patients with at least 1 MAVE was 3 of 19 patients (proportion of patients = 15.8%; 95% CI, 4.17% to 40.49%) in the prior ravulizumab cohort and 1 of 9 patients (proportion of patients = 11.1%; 95% CI, 0.58% to 49.33%) in the prior high-dose eculizumab cohort. These 4 patients discontinued crovalimab treatment before 24 weeks and, as a conservative analysis approach, were regarded as having had a MAVE. There were no patients with MAVE in the C5 polymorphism cohort.¹⁷

Summary of the Pooled Safety Data From COMMODORE 1, COMMODORE 2, and COMMODORE 3 Studies

██████ ██████ ████████ █████ ██ ████ ████ █████████ ██ █████████ ██ ███ █ ██████████ █████ ██████████ ██ ███ ████ █████████ ██ ███ ████████ ███ ███ █████ ██████████ ██████████ ██ █ ████ ███ ███ █████ ██████████ ██████████ ██ █ █████ ███ ██████████ ██████████ ███ ███████ ███████ ██ █████████ ███████ ██████████ █████████ █████████ ███████ ██████████ █████ ██ █ ███ ███ ██████████ ██████ ██ █ ████ ██ ███ ██████ ██████ █████████ ███ ██████ █████████████ ███████ ███████ ██████████ █████ ███ ██████████ ██████ █████ ██████████ ██████████████ █ ███████ ██████████ ██ ████████ ███████████ ██ █████ ███ ██ ██████ ███ ███████ █████ █ ██ ██ ██████ ███ ███████ ████ ██████ ███ ███████ █████████████████ ████ █████ ███ █████ ██ █████ █████ ███ ██ ███ ██████████ ██████ ████████████ █████████████ █ ██████ ██████████ ██ ████ ███████████ █████████████████ ███ ██████ ███ ██████ ██ ███ ██████████ █████ ████████████ ██ ███ ██████████ ██████ ███████████ ████████████ ███████████.

A pooled safety analysis from COMMODORE 1, COMMODORE 2, and a single-arm study (COMMODORE 3) was submitted by the sponsor. Across the 3 trials, a total of 11 pediatric patients (9 treatment-naive and 2 switch patients with PNH) weighing 40 kg or more were treated with crovalimab up to the CCOD. Of these, 9 (81.8%) pediatric patients reported at least 1 AE. No pediatric patients reported a fatal AE, an SAE, or an AE that led to drug interruption and/or modification or withdrawal.^10,35

Critical Appraisal

Internal Validity

Methods for randomization and allocation concealment were appropriate in both the COMMODORE 2 and COMMODORE 1 trials (i.e., a central, block-based randomization procedure was used), where randomization was stratified based on prespecified relevant prognostic factors (i.e., the LDH level and number of pRBC units administered before randomization in the COMMODORE 2 trial and transfusion history in the COMMODORE 1 trial). The clinical experts consulted for this review noted that the selection of these stratification factors was appropriate given their prognostic value for the efficacy end points, including hemolysis control and transfusion avoidance. In both trials, most baseline characteristics were similar between the randomized crovalimab and eculizumab arms. However, there were numeric differences in a few of the characteristics. In the COMMODORE 2 study, they included race (64% Asian and 33% white in the crovalimab arm; 74% Asian and 23% white in the eculizumab arm), the median PNH clone size for erythrocytes (25% in the crovalimab arm and 45% in the eculizumab arm) and for granulocytes (60% in the crovalimab arm versus 75% in the eculizumab arm), as well as the history of MDS before enrolment (4% in the crovalimab arm versus 9% in the eculizumab arm); in the COMMODORE 1 study, numeric differences in certain characteristics included the median PNH clone size for monocytes (89% in the crovalimab arm versus 96% in the eculizumab arm). Nevertheless, the clinical experts did not consider that the between-group imbalances in these characteristics would impact the efficacy and safety results in the studies. For the COMMODORE 2 study, the inclusion criteria focused on patients with clinically significant disease activity. By excluding those previously treated with complement inhibitors, the COMMODORE 2 study minimized confounding from prior therapies, supporting the study’s aim of isolating crovalimab effects.

In the COMMODORE 2 trial, the predefined NIMs for the coprimary end points in the comparison of crovalimab to eculizumab (i.e., the proportion of patients with transfusion avoidance [the NIM for 95% CI lower limit was −20%] and hemolysis control [the NIM for 95% CI lower limit of the OR was 0.2]) were determined based on data reported in protocol of the 301 study,⁶⁰ comparing eculizumab-treated patients with untreated patients from the Global PNH Patient Registry database. Based on the CDA-AMC review team’s assessment, considering the rarity of PNH and that no prior studies compared crovalimab to eculizumab, the determination of these NIMs for the efficacy end points of transfusion avoidance and hemolysis control in the COMMODORE 2 trial were considered appropriate. For the change in FACIT-F score, the NIM of −5 points was considered appropriate for a population with PNH based on the literature.¹⁴ Detailed methods were not reported regarding the selection of the NIMs for the proportion of patients with BTH (the NIM for 95% CI upper limit was 20%) and the proportion of patients with stabilized hemoglobin (the NIM for 95% CI lower limit was −20%). Concerns about the level set for the NIMs were alleviated by the fact that for all 3 outcomes, a more conservative margin would have been met. In addition, all of the NIMs in the COMMODORE 2 trial were predefined, which is appropriate.

No major protocol deviations were noted in either group in the COMMODORE 2 study. The sensitivity analyses and per-protocol analyses (these data are not included in this report) were completed and supported the conclusion of noninferiority for the coprimary end points (transfusion avoidance and hemolysis control).

The COMMODORE 1 study originally planned to recruit 200 patients to the crovalimab and eculizumab arms with a primary efficacy objective of a noninferiority assessment of mean percentage change in LDH levels from baseline to week 25, and with secondary objectives of a noninferiority assessment for end points including transfusion avoidance, BTH, hemoglobin stabilization, and fatigue. Also, it was planned that the superiority of crovalimab versus eculizumab would be evaluated provided that noninferiority had first been demonstrated for all primary and secondary end points in the COMMODORE 1 trial protocol version 1.¹⁰ However, the randomization was stopped early (in November 2022 per protocol amendment version 6) due to the introduction of ravulizumab to the treatment landscape and a reduced pool of patients treated with eculizumab over time.¹⁰ The review team and the clinical experts consulted for this review assessed that the early termination of randomization in the COMMODORE 1 trial likely did not bias the study results because the reason for termination was unrelated to the observed intervention effect (e.g., when early results show a noninferiority or statistically significant superiority effect),⁶⁷ and the baseline characteristics were generally comparable between the randomization arms. Nonetheless, with this change (randomization was terminated in November 2022), the initially targeted sample size for the randomized arms of approximately 200 patients could not be reached, providing insufficient statistical power for efficacy analyses. The results of the exploratory efficacy analyses were reported descriptively, with no formal statistical noninferiority or superiority testing, limiting the interpretation and certainty of efficacy results of the COMMODORE 1 study.

Although a between-group difference was noted in the proportions of patients who discontinued from the study before 24 weeks (4.4% in the crovalimab arm versus 1.4% in the eculizumab arm in the COMMODORE 2 study, and 0% in the crovalimab arm versus 4.5% in the eculizumab arm in the COMMODORE 1 trial), the CDA-AMC review team and the clinical experts consulted for this review did not consider that the difference would impact the study results because the numbers of events were small, and the top reasons for discontinuation were physician decision or withdrawal by study participants in the crovalimab arm (Table 12). The impact of missing data was minimal in the COMMODORE 2 study because most patients completed the randomized, 24-week, primary treatment period (95.6% in the crovalimab arm and 98.6% in the eculizumab arm). In the COMMODORE 1 study, the data of 86.7% of patients in the crovalimab arm and 79.5% of patients in the eculizumab arm who completed the randomized treatment period were available for analysis. Of note, in the COMMODORE 1 study, 5 patients in each treatment arm were ongoing in the study at the CCOD; thus, the proportions of patients with missing data due to discontinuation during the randomized period were 4.4% in the crovalimab arm and 6.8% in the eculizumab arm. The impact of missing data in the COMMODORE 1 study was considered as likely being small.

There were imbalances for some previous or concomitant treatments (Table 15). In the COMMODORE 2 study, between-group differences were observed in the proportions of patients who had corticosteroids for dermatological preparations (39% in the crovalimab arm and 67% in the eculizumab arm) and systemic use (39% in the crovalimab arm versus 65% in the eculizumab arm), which was considered unexpected by the clinical experts; in the COMMODORE 1 study, the differences were noted for antibacterials for systematic use (71% in the crovalimab arm versus 50% in the eculizumab arm), otological treatments (56% in the crovalimab arm versus 27% in the eculizumab arm), and ophthalmological and otological preparations (49% in the crovalimab arm versus 14% in the eculizumab arm). Nonetheless, the clinical experts noted that the between-group imbalance for these previous or concomitant treatments in the 2 trials likely did not impact the efficacy and safety results.

The clinical experts consulted for this review noted that the maintenance dose of eculizumab (900 mg maintenance every 2 weeks for IV infusion) in both pivotal studies aligned with its product monograph¹¹ and clinical guidelines. The clinical experts further noted that approximately one-third of patients require higher doses of eculizumab than recommended by Health Canada to properly control symptoms and signs associated with PK breakthrough. Dose modifications for eculizumab were not permitted in the COMMODORE 1 and COMMODORE 2 trials (similar to the noninferiority 301 study⁶⁰ comparing ravulizumab versus eculizumab). For the COMMODORE 2 trial, the lack of permitted dosage adjustments may have biased the efficacy results in favour of crovalimab relative to how eculizumab is dosed in clinical practice; however, the magnitude of this potential bias is unclear. The clinical experts noted that in clinical practice, patients are observed on eculizumab treatment for at least 3 months after initiation before considering a dose increase for patients in whom the 900 mg dose is considered to be inadequate. The clinical experts noted that the randomized assessment period (24 weeks) in the COMMODORE 2 trial was likely too short to explore the impact of increasing the eculizumab dose. For the COMMODORE 2 trial, there was no randomized comparative evidence of crovalimab versus eculizumab at a higher dose than that recommended by Health Canada in patients who are C5 inhibitor–naive. For the COMMODORE 1 trial, the inclusion criteria specified treatment with eculizumab according to the approved dosing recommended for PNH (900 mg maintenance every 2 weeks) and completion of a minimum of 24 weeks of treatment before day 1. It is likely that patients requiring a higher or more frequent dose of eculizumab beyond the product monograph–recommended dosage would have been excluded from the COMMODORE 1 trial. For the COMMODORE 1 study, there is no randomized evidence to inform the efficacy or safety of crovalimab in patients who switched from eculizumab given at higher doses than recommended by Health Canada.

The clinical experts consulted for this review noted that LDH is a sensitive marker for intravascular hemolysis that was assessed at multiple time points, and hemolysis control defined as an LDH value of 1.5 or less multiplied by ULN from week 5 (in the COMMODORE 2 trial) and from baseline (in the COMMODORE 1 study) through the end of the primary treatment period was appropriate. FACIT-F and EORTC QLQ-C30 GHS and QoL are validated measurements, but the CDA-AMC review team was unable to find studies about validation of EORTC IL-40 symptoms or the PPQ developed by the sponsor. Fatigue and HRQoL results were evaluable in most patients in the COMMODORE 2 trial (96%) and in 76% to 86% of the patients in the COMMODORE 1 study. For these outcomes, the impact of missing outcome data in the COMMODORE 2 study is minimal; in the COMMODORE 1 study, the impact of missing outcome data (14% in the crovalimab arm and 24% in the eculizumab arm) is unclear.

Both the COMMODORE 2 and COMMODORE 1 trials used open-label study designs, which could have potentially increased the risk of bias during outcome ascertainment. While laboratory outcomes such as LDH and hemoglobin are likely at a low risk of bias due to central measurement, the open-label design of the COMMODORE 1 and COMMODORE 2 trials may have impacted the reporting of subjective PROs, including FACIT-F, EORTC QLQ-C30 GHS and QoL, EORTC IL-40 symptoms, and patient treatment preference. The magnitude and direction of potential bias for the FACIT-F, EORTC QLQ-C30 GHS and QoL, and EORTC IL-40 outcomes were uncertain. Data about patient preference were collected among the patients who first received eculizumab and then switched to crovalimab. Most patients (84% to 96%) preferred crovalimab, with the top reasons including that crovalimab was easier and less time-consuming to administer, and was associated with fewer hospital visits compared to eculizumab. Due to the open-label design, the patient treatment preference outcome was particularly prone to bias in favour of crovalimab. Furthermore, the sponsor developed a PPQ that included a list of 13 predefined reasons for the preference, which may have introduced bias if predefined reasons did not fully capture the range of opinions or preferences across patients included in the studies.

Overall, the statistical methods used in both trials were appropriate. The COMMODORE 2 study was powered on its coprimary outcomes (hemolysis control and transfusion avoidance); however, the subgroup analyses based on key baseline demographic and disease characteristics for these 2 coprimary outcomes were likely underpowered to identify or confirm subgroup differences. For hemolysis control in the COMMODORE 2 study, a total of 7.5% of missing central LDH values were reported from week 5 through week 25 (9.5% in the crovalimab arm versus 3.7% in the eculizumab arm). The results from the respective imputation-based sensitivity analyses (MCMC algorithm, predictive mean matching algorithm, and tipping point analysis) were consistent with those from the primary analysis, supporting the robustness of the primary analysis results. For the analysis of transfusion avoidance, there were no missing data and imputation methods were not applicable.

The interpretation of results of arm C in both the COMMODORE 2 and COMMODORE 1 trials was limited due to the lack of a comparator, the limited number of patients and events, and the descriptive summary of data.

Results in the extension periods beyond week 25 (from switch baseline up to switch week 25) in the COMMODORE 2 and COMMODORE 1 trials were limited by the open-label and descriptive nature of the extension study, the lack of comparative analyses, and missing data (of patients in the arm B groups of the COMMODORE 2 and COMMODORE 1 trials — approximately 63% of patients in each trial were in the arm B switch to crovalimab cohort for at least 24 weeks before the data cut-off for analysis sets).

External Validity

Patients in the 2 pivotal trials were recruited from multiple countries. No patients were enrolled from Canada in the COMMODORE 2 study. One patient (assigned in the eculizumab arm) was from 1 site in Canada in the COMMODORE 1 study. The majority of patients enrolled in the COMMODORE 2 study were Asian and in the COMMODORE 1 study were white. There were relatively low rates of those who did not advance past screening in the COMMODORE 2 study (12%) and the COMMODORE 1 study (13%), mainly due to a lack of willingness to comply. The clinical experts did not regard them as factors that would influence the generalizability of the studies. The clinical experts noted that overall, the eligibility criteria of patients in both trials aligned with the diagnosis standard and treatment indication for PNH in clinical settings, and the demographic and disease characteristics of the patients (including LDH levels and history of transfusion) were mostly aligned with the patients seen in clinical practice in Canada. Of note, in both trials, 1 patient inclusion criterion was having a granulocyte or monocyte clone size of 10% or more. Fewer than 10 patients had a baseline PNH clone size for granulocytes of less than 10%, and no patients had a baseline PNH clone size for monocytes of less than 10%.¹⁵ In the COMMODORE 1 trial, the number of patients with a PNH clone size of less than 10% for granulocytes or monocytes was not reported, but based on the median and range values, the CDA-AMC review team estimates that the numbers of such patients are likely to be small. There may be some uncertainties in the generalizability of results from the 2 pivotal studies in patients with a PNH clone size of less than 10% for these 2 biomarkers. The clinical experts consulted for this review noted that the patients excluded from the COMMODORE 2 and COMMODORE 1 trials due to not fulfilling the inclusion criteria of having a granulocyte or monocyte clone size of 10% or more are typically asymptomatic and would less likely be considered for treatment. If these patients were showing other disease manifestations indicating a therapy, crovalimab could also be used with anticipated similar efficacy and safety results as seen in the 2 trials as per the clinical experts.

The sample size of pediatric patients was small in the COMMODORE 2 trial (n = 6) and in the COMMODORE 1 trial (n = 1), and only descriptive results were available. The clinical experts anticipated that pediatric patients would have efficacy results similar to those in the main trial arms; however, there is a need for enhanced safety consideration regarding the risk of infections, including meningitis in pediatric patients.

The recommended body weight–based doses of crovalimab once every 4 weeks via SC injection were considered as adequate and reasonable in clinical practice in Canada by the clinical experts. The clinical experts pointed out the importance of monitoring patients for any occurrence of kidney dysfunction or infections (particularly meningitis) during the treatment of crovalimab.

The trials included outcomes that were important to patients, including hemolysis control assessed by LDH relative to the ULN, stabilized hemoglobin and maintenance, transfusion outcomes, fatigue measurement, and HRQoL. These outcomes, together with other outcomes reported in this review (i.e., EORTC IL-40 selected symptoms scales) and patient preference for crovalimab or eculizumab, were considered appropriate by the clinical experts and the clinician group. The randomized period of the 2 trials was 24 weeks. According to the clinical experts and the clinician group, given that compliance may impact treatment effects, longer-term comparative evidence on the durability of the effectiveness of crovalimab would be informative. Likewise, the occurrence of some AEs, especially rare ones (e.g., meningitis), may take longer than 24 weeks to be identified. Comparative longer-term follow-up to assess safety between crovalimab and other complement inhibitors would be preferred.

There was a lack of direct comparison between crovalimab and ravulizumab in C5-naive patients and in patients who switched from ravulizumab to crovalimab. Ravulizumab and eculizumab have the same mechanism of action and ravulizumab has demonstrated similar benefits in adult patients with PNH compared with eculizumab in 2 RCTs (the 301 study and the 302 study). The clinical experts anticipated similar efficacy between crovalimab and ravulizumab and highlighted the novel route of administration and schedule for crovalimab. Based on the results from the COMMODORE 1 and COMMODORE 2 trials, the clinical experts supported initiating crovalimab or switching a patient from a currently available C5 inhibitor (eculizumab or ravulizumab) to crovalimab in patients believed to benefit from crovalimab’s SC administration and schedule (for patients who plan for pregnancy or are pregnant, eculizumab would be used).

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for outcomes considered most relevant to inform expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group:^12,13

• “High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate — The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word ‘likely’ for evidence of moderate certainty (e.g., ‘X intervention likely results in Y outcome’).

• Low certainty: Our confidence in the effect estimate is limited — The true effect may be substantially different from the estimate of the effect. We use the word ‘may’ for evidence of low certainty (e.g., ‘X intervention may result in Y outcome’).

• Very low certainty: We have very little confidence in the effect estimate — The true effect is likely to be substantially different from the estimate of effect. We describe evidence of very low certainty as ‘very uncertain.’”

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, the imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty of evidence assessment was the presence or absence of an important effect based on threshold (MID) in the literature for the FACIT-F score in both trials. For the COMMODORE 2 study, the target of the certainty of evidence assessment was the presence or absence of a predefined NIM per study protocol for hemolysis control, BTH, stabilized hemoglobin, having reached and maintained a minimum hemoglobin level, and transfusion avoidance. Due to the lack of a formal MID estimate or NIM value, the target of the certainty of evidence assessment was the presence or absence of any (non-null) effect for mean percentage change in LDH levels, pRBC units transfused, EORTC QLQ-C30 GHS and QoL score, infections, and deaths. For the COMMODORE 1 study, due to the lack of a formal MID estimate or NIM (after amendment of the COMMODORE 1 study, all the efficacy end points became exploratory and descriptive; thus, NIM values were not applicable), the target of the certainty of evidence assessment was the presence or absence of any (non-null) effect for hemolysis control, BTH, mean percentage change in LDH levels, stabilized hemoglobin, having reached and maintained a minimum hemoglobin level, transfusion avoidance, pRBC units transfused, EORTC QLQ-C30 GHS and QoL score, infections, and deaths.

For the GRADE assessments, the COMMODORE 2 and COMMODORE 1 studies were assessed individually because the COMMODORE 2 study included patients naive to complement inhibition and the COMMODORE 1 study included complement inhibition–experienced patients.

Results of GRADE Assessments

Table 2 and Table 3 present the GRADE summary of findings for crovalimab versus eculizumab for patients with PNH who were naive to and had experienced complement inhibitors, respectively.

Long-Term Extension Studies

Content in this section has been informed by materials submitted by the sponsor. The following has been summarized and validated by the review team.

Description of Studies

The phase I and phase II COMPOSER trial (NCT03157635) consisted of 4 sequential parts and an OLE. This section includes a summary of the OLE period that evaluated the safety, tolerability, PKs, and PDs of crovalimab in patients with PNH aged 18 years to 75 years who were treatment-naive or who switched from eculizumab. Participants were enrolled at 14 sites in 6 countries: Germany (3 sites), Japan (5 sites), France (1 site), Hungary (2 sites), Korea (2 sites), and Italy (1 site). The study did not include sites in Canada.

Part 1 of the COMPOSER trial was designed to evaluate the safety, tolerability, PKs, and PDs of crovalimab in healthy volunteers and part 2, part 3, part 4, and the OLE period evaluated safety, tolerability, PKs, and PDs of crovalimab in patients with PNH. Part 1 evaluated the safety and tolerability of single doses of crovalimab in healthy volunteers and assessed PD and PK outcomes. Methods and results for part 1 are not further summarized in this CDA-AMC Clinical Review Report. Part 2, part 3, and part 4 comprised a dose-escalation study (planned to enrol 6 patients), a multidose, single-arm study (planned to enrol 18 patients), and a multidose, 2-arm study (planned to enrol 10 patients) in patients with PNH, respectively. Results for part 2, part 3, and part 4 are not further described in this CDA-AMC Clinical Review Report.

The OLE was initiated once the first patient in part 2 was eligible for the OLE. Patients transitioning from part 2, part 3, or part 4 initially remained on the same SC dose regimen as they were receiving in part 2, part 3, and part 4, respectively.

Populations

This is an OLE for patients who participated in part 2, part 3, and part 4 and who, in the opinion of the investigators, derived benefit from treatment with crovalimab. Part 2, part 3, and part 4 of the COMPOSER study enrolled patients with PNH of a population similar to that of the pivotal trials.

Interventions

The OLE phase was available for patients after completing 20 weeks of part 2, part 3, or part 4. Patients enrolling in the OLE period from part 2, part 3, and part 4 initially stayed on the previously assigned treatment schedule. Following analysis of the primary treatment period for part 2, part 3, and part 4, weight-based tiered dosing was introduced in the OLE phase to ensure that all patients received a comparable drug exposure across the body weight continuum. With the implementation of the tiered weight dosing, all patients in the OLE phase who were not currently receiving dosing every 4 weeks were required to switch to the dosing every 4 weeks regimen. Patients either received 680 mg SC every 4 weeks (body weight ≥ 40 kg to < 100 kg) or 1,020 mg SC every 4 weeks (body weight ≥ 100 kg).

Dose modification was required in patients whose body weight changed by 10% or more (compared with screening or the visit when the latest dose modification occurred, whichever was later) to exceed or become equal to 100 kg or to fall below 100 kg during the course of therapy.

Patients with 2 or more qualifying intravascular hemolysis events in 24 weeks were considered for an increased crovalimab maintenance dose with approval from the medical monitor. Qualifying intravascular hemolysis events were defined as symptoms of intravascular hemolysis that occurred within a week before the next maintenance dose without an identifiable trigger. If approved by the medical monitor, patients with a body weight of 40 kg to less than 100 kg would increase their maintenance dose from 680 mg every 4 weeks to 1,020 mg every 4 weeks; patients with a body weight of 100 kg or more would increase their maintenance dose from 1,020 mg every 4 weeks to 1,360 mg every 4 weeks.

If, based on investigators’ assessment, the patient or caregiver was able to self-administer or administer crovalimab-vial SC at home, after successful SC self-administration or the administration of crovalimab-vial training, the patient or caregiver received permission by the investigators to self-administer or administer a crovalimab vial at home.

Although 1 or more additional IV doses of crovalimab could be administered as a rescue dose (340 mg IV regardless of body weight infused over 30 minutes), none were administered during the COMPOSER study.

Outcomes

The efficacy results in the OLE period (i.e., from week 20 to the CCOD of November 1, 2021) are presented in text as a pooled total study population from part 2, part 3, and part 4. At the time of enrolment in the OLE period, all patients had been on stable crovalimab maintenance during the 20-week primary treatment period and were therefore considered comparable from a safety and efficacy standpoint, irrespective of naive or switch status at the time of enrolment in the study.

The exploratory long-term efficacy end points that were analyzed in the OLE period included the following:

• hemolysis control (change in LDH) reported from week 20 to the CCOD using the mean normalized LDH level by visit and the proportion of patients by visit achieving an LDH value of 1.5 or less multiplied by ULN

• the proportion of patients achieving transfusion avoidance by 24-week intervals from week 20 to the CCOD

• the proportion of patients with stabilized hemoglobin levels by 24-week intervals from week 20 to the CCOD

• the proportion of patients with BTH events by 24-week intervals from week 20 to the CCOD

• the rate of BTH events from week 20 to the CCOD adjusted for patient-years at risk.

Safety data included AEs, reasons for withdrawal from the study, laboratory data, ECGs, concomitant medications, vital signs, and physical examination results. AESIs for this study included abnormal liver function tests, type III immune complex reactions, and suspected transmission of an infectious drug by the study drug.

Statistical Analysis

For part 2, part 3, and part 4 of the study, the efficacy analysis population was the same as the safety analysis population.

For the OLE period, the efficacy-evaluable population included all efficacy-evaluable patients who completed the primary treatment period in part 2, part 3, or part 4 of the study and were enrolled in the OLE period (18 treatment-naive patients and 25 switched patients).

The safety-evaluable population was defined as all healthy volunteers or patients who received at least 1 administration of the study drug (crovalimab or placebo) and included 18 treatment-naive patients and 26 switched patients.

At the time of the CCOD for the Update Clinical Study Report, patients had a maximum of approximately 4 years of follow-up in the OLE period. Hemolysis control (change in LDH levels) was analyzed by visit. To support pooled analysis of LDH by visit, time windows were defined to identify corresponding visits across the study parts, which have nonidentical visit schedules.

The proportion of patients with transfusion avoidance, hemoglobin stabilization, and BTH events was analyzed over sequential 24-week intervals, starting at week 20, to evaluate whether the efficacy of crovalimab, as measured by these end points, remained stable over time.

The analysis of a specific 24-week interval only included patients who had completed the entire 24-week interval at the time of the CCOD (November 1, 2021). Additionally, as the occurrence of BTH events is expected to be rare, the rate of BTH was also calculated over the entire OLE period and adjusted for patient-years at risk to take into account the individual patients’ exposure.

Due to different long-term follow-up times per patient at the CCOD, the data for some visits or intervals were based on small sample sizes (i.e., n < 10), which results in wide 95% CIs and significant uncertainty in the evidence. Therefore, for the purpose of the interpretation of the results obtained from the exploratory efficacy end points, only visits or intervals where the sample size was n equals 10 or more were considered in the following discussion.

Results

Patient Disposition

Overall, 56 patients with PNH were screened. In total, 44 patients with PNH were enrolled in part 2 (N = 10), part 3 (N = 19), and part 4 (N = 15). Of the 44 patients,18 individuals were treatment-naive and 26 patients switched from eculizumab. One patient in part 3 of the study elected not to proceed to the OLE phase. As of the CCOD (November 1, 2021), 38 patients were ongoing in the OLE period (17 treatment-naive patients and 21 patients who switched from eculizumab). Overall, a total of 5 patients were discontinued from the OLE phase (4 patients originally enrolled in part 3 and 1 patient originally enrolled in part 4, arm A) (Table 20).

The efficacy-evaluable population for the OLE period consisted of 42 patients (treatment-naive, n = 18; switched, n = 25); 1 patient in PART 3 completed the primary treatment period but elected not to enrol in the OLE phase (Table 21).

Table 20: Patient Disposition From the COMPOSER Study

FAS = full analysis set; OLE = open-label extension.

Note: One patient in part 3 completed the primary treatment period but did not enrol in the OLE period; this is not captured as a study discontinuation.

Sources: Sponsor’s submission.²³

Table 21: Summary of Baseline Characteristics From the COMPOSER Study

LDH = lactate dehydrogenase; PNH = paroxysmal nocturnal hemoglobinuria; RBC = red blood cell; SD = standard deviation; ULN = upper limit of normal.

^aTransfusion events that occurred within a year before the first dose.

Sources: Sponsor’s submission.²³

Exposure to Study Treatments

Treatment-Naive Patients With PNH

From baseline to the November 1, 2021, CCOD, treatment-naive patients with PNH received crovalimab treatment for a median duration of 3.40 years (range, 0.9 year to 4.4 years). The median number of doses was 83.5 (range, 19 to 163).

Patients With PNH Switching From Eculizumab

From baseline to the November 1, 2021, CCOD, patients with PNH switching from eculizumab received crovalimab treatment for a median duration of 2.88 years (range, 0.4 year to 3.9 years). The median number of doses was 48.5 (range, 11 to 129) (Table 22).

Table 22: Summary of Treatment Exposure (Safety-Evaluable Population, Part 2, Part 3, and Part 4 of the COMPOSER Study, as of CCOD of November 1, 2021)

CCOD = clinical cut-off date; SC = subcutaneous; SD = standard deviation.

^aPatients who switched from eculizumab.

Sources: Sponsor’s submission.²³

Concomitant Medications and Cointerventions

In addition to the study treatment, participants in part 2, part 3, and part 4 were permitted to continue receiving immunosuppressant therapy (with the exception of azathioprine and IV immunoglobulin), corticosteroids, iron supplements, and folic acid (with the exception of erythrocyte-stimulating drugs such as erythropoietin).

During the OLE phase, in addition to the study treatment, patients were permitted to concomitantly continue to receive immunosuppressant therapy (with the exception of IV immunoglobulin and azathioprine), corticosteroids, iron supplements, and folic acid (with the exception of erythrocyte-stimulating drugs such as erythropoietin). For patients in the OLE phase, revaccination considered as necessary was based on standard of care or local guidance.

Efficacy

Hemolysis Control

The proportion of patients reaching an LDH level of 1.5 or less multiplied by ULN per visit suggested there was little to no difference across visits during the OLE period, with 80% to 100% of the evaluable patients at each visit having an LDH level of 1.5 or less multiplied by ULN.

The mean normalized LDH level was generally maintained below 1.5 multiplied by ULN during the OLE phase. The point estimate of the mean normalized LDH level by visit ranged from 1.09 to 1.24 multiplied by ULN.

Transfusion Avoidance

The proportion of patients with transfusion avoidance remained relatively stable over time during the OLE period, with 82.9% to 91.7% of all patients avoiding transfusion across the 24-week intervals. Seven of the 11 patients who received a blood transfusion during the OLE period reported a history of aplastic anemia and/or anemia.

Stabilized Hemoglobin

The proportion of patients reaching hemoglobin stabilization also remained relatively stable over time during the OLE phase, with 79.5% to 87.5% of all patients reaching hemoglobin stabilization across the 24-week intervals.

Breakthrough Hemolysis

The proportion of patients with BTH events was low during the OLE period, with 0% to 4.9% of all patients reporting a BTH event across the 24-week intervals. One patient who had a BTH event discontinued the study before completing the full interval from week 20 to week 44 in which the BTH event occurred. Consequently, this BTH event was reported and included in the computation of the rate of BTH events adjusted for patient-years at risk but is not included in the calculation of the proportion of patients with BTH events per 24-week interval analyses.

Harms

From baseline to the CCOD (November 1, 2021), a total of 42 (95.5%) patients experienced at least 1 AE. There were no deaths and no AEs that resulted in withdrawal from the study. A total of 14 (31.8%) patients experienced at least 1 SAE over a median treatment duration of 3.03 years.

AEs of any grade were reported in 17 (94.4%) treatment-naive patients and in 25 (96.2%) patients who switched from eculizumab. The most frequently reported AEs were nasopharyngitis (29.5%), upper respiratory tract infection (22.7%), and pyrexia (22.7%).

Fourteen (31.8%) patients experienced SAEs — 6 (33.3%) treatment-naive patients and 8 (30.8%) patients who switched from eculizumab. In the treatment-naive patients, all SAEs occurred in single patients. In the patients who switched from eculizumab, the SAE that occurred in 2 or more patients was BTH (7.7%). All other SAEs occurred in single patients. In the treatment-naive group, 1 (5.6%) patient experienced an SAE of BTH considered by the investigators to be related to study treatment. In the switch group, 1 (3.8%) patient experienced an SAE of upper respiratory tract infection assessed as related to study treatment by the investigators. No patients experienced SAEs that led to dose modification. One treatment-naive patient experienced an SAE of cardiac failure that led to dose interruption; it was deemed not related to study treatment as assessed by the investigators.

From baseline to the November 1, 2021, CCOD, 2 AESIs of abnormal liver function tests were reported in 2 (4.5%) patients with PNH (1 treatment-naive patient [increased alanine aminotransferase] and 1 patient who switched from eculizumab [a drug-induced liver injury]). In both cases, the events were of moderate severity and not considered related to study treatment as assessed by the investigators.

Of the 26 patients who switched from eculizumab, 2 (7.7%) patients experienced AESIs of type III immune complex reactions. Both of these AESIs were mild to moderate in severity and occurred during the primary treatment period. No additional type III immune complex reactions occurred during the OLE period. One patient in the treatment-naive group experienced an AESI of type III immune complex reactions at the time of switching from crovalimab to eculizumab, after discontinuing from the study due to a lack of efficacy. This AESI was mild in severity.

Other than type III immune complex reactions, no other hypersensitivity events were reported in patients with PNH during this study up to the November 1, 2021, CCOD. In addition, no AESIs of suspected transmission of an infectious drug by the study drug or cases of meningococcal meningitis were reported.

A total of 2 (4.5%) patients experienced at least 1 injection site reaction up to the CCOD (1 treatment-naive patient and 1 patient who switched from eculizumab). Three (11.5%) patients who switched from eculizumab reported 5 events of infusion-related reaction during the primary treatment period. The infusion-related reaction that occurred in 2 or more patients was paresthesia. No treatment-naive patients reported infusion-related reactions.

A total of 36 (81.8%) patients (77.8% of treatment-naive patients and 84.6% of switched patients) experienced at least 1 infection up to the CCOD.

There were 3 pregnancies in partners of patients reported in the study, with no reported detectable detrimental consequences to the infants (Table 23).

Table 23: Summary of Harms Results From the Long-Term Extension Study of the COMPOSER Study