Input From Clinical Experts Consulted by CADTH

The following input was provided by 1 clinical specialist with expertise in the diagnosis and management of PNH.

One unmet need of patients with PNH is that the quality of life of patients being treated with eculizumab could be improved by modifying the treatment schedule or ease of treatment administration. Additionally, at the dosage recommended by the Health Canada–approved product monograph for eculizumab in PNH, approximately 20% of patients do not have complete control of signs and symptoms. This is associated with incomplete pharmacologic C5 inhibition and could be addressed by administering higher doses of eculizumab. Finally, patients who have clinically significant anemia secondary to eculizumab treatment have an unmet need for extravascular hemolysis control; however, extravascular hemolysis would not be expected to improve with ravulizumab treatment.

Ravulizumab has the same mechanism of action as eculizumab and if funded, would be considered first-line therapy in place of eculizumab for most patients. Patients in need of anti-complement therapy include those with evidence of a PNH clone (usually white blood cell clone size > 10%), hemolysis (i.e., LDH > 1.5 × the upper limit of normal [ULN]), and symptoms. Almost all, if not all, patients with hemolytic PNH who would qualify for eculizumab would similarly be expected to respond to ravulizumab. Neither treatment would be effective in the small proportion of patients of Japanese (approximately 3%) and Han-Chinese (approximately 1%) descent who have a polymorphism which negates the effect of eculizumab as there is no effective target on C5. Currently, eculizumab may be preferred over ravulizumab during pregnancy given the available efficacy and safety data for eculizumab although this may change as more clinical experience with ravulizumab accumulates.

A clinically meaningful response to treatment would include improved symptoms and signs (e.g., fatigue, dyspnea, kidney function, abdominal pain, erectile dysfunction) and/or reduced transfusion demands. Response is assessed by review of the signs and symptoms and mapping onto biochemical evidence of reduced intravascular hemolysis (LDH < 1.5 × ULN) and improved blood counts (e.g., hemoglobin), and other parameters (e.g., creatinine, echocardiogram). Discontinuation of anti-complement therapy would rarely be considered and relevant situations would include: nonresponse (almost always associated with the polymorphism that negates the effect of C5 inhibition); persistent, severe adverse reactions (very rare); progression to severe bone marrow failure requiring bone marrow transplant; and regression of PNH clone to less than 10%, if associated with resolution of clinically significant hemolysis.

Clinician Group Input

Clinician group submissions were not received for this review.

Description of Studies

Two relevant studies, the ALXN1210-PNH-301 and ALXN1210-PNH-302 studies (referred to here as Study 301 and Study 302, respectively), were selected for inclusion in the CADTH systematic review. Both studies were open-label, active-controlled, parallel-group, noninferiority, randomized controlled trials (RCTs). Both studies were sponsored by Alexion Pharmaceuticals, Corp. and the primary evaluation periods of both studies took place from 2016 to 2018. Study 301 (N = 246) enrolled adult patients with PNH who were treatment-naive, whereas Study 302 (N = 197) enrolled adult patients with PNH who had been receiving eculizumab. Patients were randomized 1:1 to ravulizumab or eculizumab. Noninferiority of ravulizumab compared with eculizumab was assessed for transfusion avoidance, fatigue, breakthrough hemolysis, LDH normalization, and hemoglobin stabilization during a 26-week primary evaluation period.

Patients in both studies were required to have a PNH diagnosis confirmed by flow cytometry (granulocyte or monocyte clone size of at least 5%) and patients in Study 301 were required to have an LDH level of at least 1.5 × ULN and at least 1 PNH-related sign or symptom in the past 3 months. Patients in Study 302 were required to have received eculizumab and have controlled LDH for at least the 6 months before the study. Across both studies, approximately half of patients were male, most were either Asian or White, and mean age was 45 years to 49 years. In Study 301, most patients had an LDH level of 3 × ULN or greater and had received at least 1 transfusion in the past year. Patients in Study 302 had a mean LDH of 228 U/L to 235 U/L (with the ULN for LDH considered to be 246 U/L), with 12.2% to 13.4% of patients having received at least 1 transfusion in the past year. Patients in Study 301 had a shorter mean disease duration (6.4 years to 6.7 years) than patients in Study 302 (11.9 years to 12.4 years), who had been receiving eculizumab for a mean of 5.6 years to 6.0 years. There were lower percentages of patients in Study 301 who had experienced a major adverse vascular event (MAVE; 13.6% to 20.7%) than in Study 302 (22.4% to 28.9%).

Efficacy Results

The results for transfusion avoidance, LDH normalization, and percentage change in LDH level, which were the primary and coprimary end points in the studies, are presented in Table 2. Results for other clinically important outcomes, health-related quality of life (HRQoL) and fatigue, are also presented. The results for the per-protocol (PP) analyses for all primary and key secondary end points were consistent with the primary analyses.

Harms Results

The results for adverse events (AEs) are presented in Table 2. Most patients (86.8% to 88.0%) in both treatment groups in both studies reported at least 1 AE. The most common AE was headache and there were no notable imbalances in AEs. Serious AEs (SAEs) were reported in 4.1% to 8.8% of each treatment group in both studies. The most common SAEs were hemolysis and pyrexia, which occurred in 3.1% or less of each treatment group. There were no withdrawals due to AEs in either study. One patient in the eculizumab group in Study 301 died due to lung adenocarcinoma during the extension phase of the study.

In terms of notable harms, serious infections were reported in 1.0% to 3.3% of each treatment group in both studies. Infusion reactions were reported in 3.1% to 8.8% of patients across each treatment group in both studies. As for treatment-emergent anti-drug antibody (ADA)–positive samples, there was 1 in each treatment group in Study 301 and 1 in the eculizumab group in Study 302 and titres were considered to be low.

Table 2

Summary of Key Results From Pivotal and Protocol Selected Studies.

Critical Appraisal

The pre-specified noninferiority margins for the primary and key secondary end points (aside from percent change in LDH, potentially) were based on a magnitude of loss of benefit that may not be clinically acceptable. However, there are several factors that mitigate the risk of unacceptable loss of benefit with ravulizumab versus eculizumab. These include that all of the primary and key secondary end points met their respective noninferiority margins, there were minimal missing data, the PP analyses were consistent with the primary analyses for all end points, and a more conservative margin would have been met for all end points.

The open-label nature of the studies means that outcomes relying on subjective reporting, such as the EORTC QLQ-C30 and the FACIT-F, could have been biased with potential for bias in favour of ravulizumab. Additionally, the reliability, validity, and responsiveness of the EORTC QLQ-C30 and the FACIT-F have yet to be characterized in patients with PNH. Statistical testing was only performed for the FACIT-F score and not for other symptom assessments or for the EORTC QLQ-C30 scales.

The criteria for Study 302 were chosen in such a way that patients requiring a higher dose or more frequent dosing of eculizumab beyond the product monograph-recommended dosage would have been excluded. While these patients were included in Study 301, the studies did not allow for deviation from the labelled dosage of eculizumab (900 mg maintenance dose) and this may have biased the efficacy results in favour of ravulizumab relative to how eculizumab is dosed in clinical practice.

Description of Studies

Safety and efficacy results from the respective extension periods for Study 301 (N = 243) and Study 302 (N = 191), during which all patients received ravulizumab, were also submitted by the sponsor and are presented in this report for the 26-week period following the randomized treatment period. Also included in the sponsor’s submission was a patient preference substudy (N = 95) which allowed patients to enrol from Study 302 who enrolled in the extension period and had received at least 2 doses of ravulizumab during the extension period. A novel patient preference questionnaire was developed for the study and the objective of the study was to assess patient preferences for ravulizumab or eculizumab and to identify the key factors influencing preference.

Efficacy Results

The results from the extension periods of Study 301 and Study 302 were reported as summary statistics and indicated that efficacy as assessed through transfusion avoidance, FACIT-F score, breakthrough hemolysis, LDH normalization, and hemoglobin stabilization was generally maintained with ravulizumab treatment for another 26 weeks following the randomized treatment period.

According to the results from the questionnaire administered in the patient preference substudy, 93% of patients preferred ravulizumab overall with 43% of patients choosing frequency of infusions and 23% of patients choosing overall quality of life as the most important treatment factor when deciding preference.

Harms Results

The AE profiles in the extension periods of Study 301 and Study 302 were similar to those in the randomized treatment periods, with no new safety signals identified. The frequency of headaches numerically decreased between the 2 periods in both treatment groups in both studies.

Critical Appraisal

The extension periods of Study 301 and Study 302 do not provide evidence for the comparative efficacy of ravulizumab versus eculizumab because all patients who continued in the extension periods received ravulizumab. As well, reductions in sample size in periods beyond the first 52 weeks of study treatment precluded the ability to assess results beyond 1 year of treatment, which is a concern given the chronic nature of the disease.

There were several limitations identified in the patient preference substudy that introduce substantial uncertainty in the results. These include the lack of evidence for the reliability and responsiveness of the questionnaire, the potential for recall bias given that ravulizumab was the most recent treatment for all patients, the small sample size relative to the population of Study 302, and uncertainty surrounding reasons for the reduction in sample size.

Unmet Needs

Not all patients with PNH have complete control of intravascular hemolysis with eculizumab. Approximately 20% of patients require higher doses of eculizumab or more frequent doses than recommended in the Health Canada–approved product monograph for the PNH indication to properly control symptoms and signs associated with pharmacokinetic breakthrough (i.e., insufficient pharmacologic C5 inhibition). Higher doses (i.e., the recommended 1,200 mg maintenance dose for other indications) or more catered doses (e.g., weight-based doses) are needed to more fully suppress C5.

In patients receiving eculizumab, quality of life and impact on patients’ lives could be improved by modifying the treatment schedule or ease of treatment administration. Some examples would be less frequent infusions than every 2 weeks, self-administered subcutaneous injections, or oral therapies.

Extravascular hemolysis control remains an unmet need for those who have clinically significant anemia secondary to eculizumab treatment, which is caused by C5 blockade and would not be expected to improve with ravulizumab treatment. Proximal complement inhibitors have the potential to minimize this complication in PNH, with pegcetacoplan having received FDA approval and various trials currently under way. However, these are not clinically available yet in Canada and C5 blockade is still of utmost importance to control intravascular hemolysis.

Place in Therapy

Ravulizumab has the same mechanism of action as eculizumab. Its modifications allow more profound C5 blockade and greater recycling of the molecule, thereby extending the treatment cycle from every 2 weeks to every 8 weeks. Although ravulizumab is not the first treatment to address the disease process, it is expected to improve some aspects of treatment, including pharmacokinetic breakthrough risk and patient convenience. If funded, ravulizumab would be considered first-line therapy in place of eculizumab for most patients and other supportive therapies would remain the same. Patients already being treated with eculizumab can be directly transitioned from eculizumab to ravulizumab.

Patient Population

PNH is an ultra-rare disease (annual incidence of approximately 5 to 10 cases per million persons) and manifests usually in vague and general or common symptoms (e.g., fatigue, thrombosis, iron deficiency anemia). As such, it can take years for the diagnosis to be made (usually following referral to a hematologist). Testing is made by high-sensitivity flow cytometry, which is available in most, if not all, academic centres. Some community labs have agreements to send samples to academic labs for processing. Along with PNH flow cytometry, the identification of intravascular hemolysis (with elevated LDH, undetectable haptoglobin, and negative direct antiglobulin test) is fairly straightforward as long as the physician thinks to test for it.

Patients without evidence of hemolysis with small PNH clones (usually < 10%, and more often < 1%) would not usually require treatment with complement blockade. Their PNH clones would be most commonly associated with a diagnosis of aplastic anemia or myelodysplastic syndrome.

Patients in need of anti-complement therapy include those with evidence of a PNH clone (usually white blood cell clone size > 10%), hemolysis (i.e., LDH > 1.5 x ULN), and symptoms. The current criteria for reimbursement of eculizumab in Canadian jurisdictions, which have not changed since eculizumab’s approval in 2009, stipulate a diagnosis of PNH based on PNH clone size, LDH level, and at least 1 of the following signs or symptoms of PNH: a thrombotic or embolic event, at least 4 units of RBCs transfused in the previous 12 months, chronic or recurrent anemia where causes other than hemolysis have been excluded, pulmonary insufficiency, renal insufficiency, or smooth muscle spasm. The severity of the criteria surrounding signs or symptoms are challenging, particularly the smooth muscle spasm criterion which requires hospitalization and/or use of narcotic analgesia. Other symptoms should also be considered as these are all signs of systemic complement dysregulation and intravascular hemolysis: thrombosis, anemia, dyspnea, pulmonary hypertension, kidney failure, abdominal pain, dysphagia, refractory erectile dysfunction, and persistent fatigue. Adding refractory erectile dysfunction and persistent fatigue to the list of signs and symptoms specifically could be beneficial for patient access. Another issue with the current reimbursement criteria is that most jurisdictions only consider granulocytes for PNH clone (granulocyte clone size > 10%) when white blood cells clone size greater than 10% would be a more appropriate criterion. Although uncommon, some patients with PNH have a monocyte clone size of greater than 10% and a neutrophil clone size of less than 10%.

Almost all, if not all, patients with hemolytic PNH who would qualify for eculizumab would similarly be expected to respond to ravulizumab. Neither treatment would be effective in the small proportion of patients of Japanese (approximately 3%) and Han-Chinese (approximately 1%) descent who have a polymorphism which negates the effect of eculizumab as there is no effective target on C5. As such, ravulizumab would be a replacement for eculizumab in most situations. One area where current perspective suggests eculizumab is preferred is in the context of pregnancy given the currently available efficacy and safety data.¹¹ Patients who are pregnant would be treated with eculizumab during pregnancy and perhaps for 4 weeks to 6 weeks postpartum; however, before and after, they could be treated with ravulizumab. This may change as more clinical experience with ravulizumab accumulates.

Assessing Response to Treatment

The following outcomes are used in clinical trials and in clinical practice: LDH normalization (LDH < 1.5 × ULN) as a surrogate for hemolysis (associated with a reduction in risk of thrombosis and other symptoms), transfusion independence in patients who required transfusions before treatment, hemoglobin stabilization (in concert with LDH normalization), improvement in PNH symptoms and signs, absence of thrombosis, and improvement in quality of life. A clinically meaningful response to treatment would include improved symptoms and signs (e.g., fatigue, dyspnea, kidney function, abdominal pain, erectile dysfunction) and/or reduced transfusion demands. There is evidence for improved survival with eculizumab treatment and the same is also anticipated for ravulizumab.

Patients with evidence of hemolysis with LDH are the easiest to monitor for a response as there would be rapid LDH reduction after the drug is started (within 4 weeks in most cases, if not faster). No patients have been identified who would not respond (unless it is known that they hold the C5 polymorphism described above). Suspicion for this would come when there is no improvement in LDH and symptoms after starting treatment and in the right ethnic context (Japanese or Han-Chinese).

Patients are typically followed every 2 weeks to 4 weeks at the time of treatment initiation to follow symptoms, laboratory evidence of improvement (e.g., LDH reduction), and to monitor for safety signals or AEs. Response is anticipated within the first 2 months of starting therapy. Patients are then monitored approximately every 3 months to 6 months once they are felt to be stable with fully suppressed intravascular hemolysis and terminal complement activity. Response is assessed by review of the signs and symptoms and mapping onto biochemical evidence of reduced intravascular hemolysis (LDH < 1.5 × ULN) and improved blood counts (e.g., hemoglobin), and other parameters (e.g., creatinine, echocardiogram, and so on). Patients are also monitored with PNH flow cytometry every 6 months to 12 months for changes in their clone size.

Discontinuing Treatment

There are a few situations where discontinuation of complement blockade would be considered:

• instances of nonresponse, which are very rare and almost always associated with the C5 polymorphism that negates the effect of C5 inhibition
• instances of persistent, severe adverse reactions, which are also very rare
• progression to severe marrow failure that would require bone marrow transplant
• regression of the PNH clone to less than 10% which would usually correspond to resolved or resolving hemolysis

Prescribing Conditions

Hematologists typically diagnose, treat, and monitor PNH patients in Canada. As the disease is extremely rare, referral to Canadian PNH Network Centres is encouraged, and these centres assume care of these patients or offer a shared-care model with the community physician.

Treatment is usually given in the community, either infused in an infusion clinic or at the patient’s home. A hospital setting is uncommon for treatment except in situations where admission is otherwise required (e.g., elective surgery). In those situations, it is desirable to have the drug available in the event of an acute exacerbation.

Inclusion and Exclusion Criteria

Details on key inclusion and exclusion criteria for the studies are presented in Table 6. Patients in both studies had to have a PNH diagnosis confirmed by flow cytometry (granulocyte or monocyte clone size of at least 5%). Patients in Study 301, the treatment-naive population, had to have an LDH level of 1.5 × ULN or greater and at least 1 of the following PNH-related signs or symptoms in the past 3 months: fatigue, hemoglobinuria, abdominal pain, dyspnea, anemia (hemoglobin < 10 g/dL), history of a MAVE, dysphagia, erectile dysfunction, or history of packed RBC transfusion due to PNH. Patients in Study 302, the eculizumab-treated population, had to have received eculizumab according to the labelled dosing recommendation for PNH and have controlled LDH (< 2 × ULN) and no MAVE for at least 6 months and a screening LDH of 1.5 × ULN or greater. According to the clinical expert consulted by CADTH for the review, these criteria would have been sufficient to exclude patients who would receive an eculizumab dosage beyond the Health Canada–approved PNH dosage due to consistent pharmacokinetic-related breakthrough hemolysis. In both studies, patients with platelet count of less than 30,000/mm³ or absolute neutrophil count of less than 500/µL were excluded. The clinical expert noted that these criteria may have been implemented to exclude patients with frank bone marrow failure, though patients with these counts in clinical practice can still receive treatment if they otherwise have evidence of hemolytic PNH.

Baseline Characteristics

Across Study 301 and Study 302, approximately half of patients were male, most were either Asian or White, and mean age was 45 years to 49 years. Although the distribution of patients among races in Study 301 was not reflective of Canadian patients and was not balanced between the ravulizumab and eculizumab groups, the clinical expert consulted by CADTH confirmed that differences in race would not notably impact outcomes. In Study 301, most patients (85.6% to 86.8%) had an LDH level of 3 × ULN or greater and had received at least 1 transfusion in the past year (82.4% to 82.6%). Patients in Study 302 had a mean LDH of 228 U/L to 235 U/L (with ULN for LDH considered to be 246 U/L), with 12.2% to 13.4% of patients having received at least 1 transfusion in the past year. Patients in Study 301 had a shorter mean disease duration (6.4 years to 6.7 years) than patients in Study 302 (11.9 years to 12.4 years), who had been receiving eculizumab for a mean of 5.6 years to 6.0 years. There were lower percentages of patients in Study 301 who had experienced a MAVE (13.6% to 20.7%) than in Study 302 (22.4% to 28.9%). Percentages of patients with a history of anemia and hematuria or hemoglobinuria were notably higher in Study 301 compared with Study 302. The clinical expert consulted by CADTH considered the percentages in Study 301 to reflect patients in the international PNH registry while a possible explanation of the lower percentages in Study 302 is recall bias due to proximity to experience with untreated disease. Although there were some imbalances between groups within each trial, they were not expected by the clinical expert to contribute any bias to the efficacy or safety results.

Table 7

Summary of Baseline Characteristics.

Thrombotic Events

In both studies, the occurrence of MAVEs during the primary evaluation period was assessed as an exploratory end point, along with the method of diagnosis, date of diagnosis, and date of resolution.

Health-Related Quality of Life

The EORTC QLQ-C30, a multidimensional, cancer-specific, self-administered questionnaire for assessing HRQoL, was administered in both studies at screening and on days 1, 8, 29, 71, 127, and 183 (or at the early termination visit if applicable). The questionnaire includes a global health status score and change in the global health status score from baseline to week 26 was a secondary end point in both studies. Scores range from 0 to 100 and a higher global health status score indicates better HRQoL. Estimates of the minimal important difference (MID) for the scale scores in patients with cancer are 10 points to 20 points for moderate changes and greater than 20 points for large change. Estimates of MIDs for the scale scores were not identified in patients with PNH. More details on the EORTC QLQ-C30 and its properties can be found in Appendix 3.

Transfusions

Sample collection for laboratory assessments occurred at each study visit and at visits where study treatment was administered, samples were collected before the dose and not from a heparinized line. Hemoglobin was assessed at screening and at each study visit (weekly until day 29, followed by visits every 2 weeks). In both studies, patients were to receive a packed RBC transfusion when they met either of the following criteria: hemoglobin value of 9 g/dL or less with signs or symptoms of sufficient severity to warrant a transfusion, or hemoglobin value of 7 g/dL or less regardless of presence of clinical signs or symptoms. The number of units to be transfused was determined by the investigator and the transfusion was recommended to take place within 48 hours of the hemoglobin determination. Patients received a transfusion before randomization and within 5 days before day 1 if they met either of the transfusion criteria. The proportion of patients achieving transfusion avoidance from day 1 through day 183 was a coprimary end point in Study 301 and a key secondary end point in Study 302. The total number of units transfused was a secondary end point in both studies.

Symptoms of PNH

Also in both studies, the FACIT-F was administered to patients at screening and on days 1, 8, 29, 71, 127, and 183 (or at the early termination visit if applicable). Percent change in FACIT-F score from baseline to week 26 was a key secondary end point in both studies. The FACIT-F scale is a 13-item questionnaire used to assess patient fatigue and energy levels. Scores range from 0 to 52, with lower scores indicating greater fatigue. Estimates of the MID for the FACIT-F score were not identified in patients with PNH. More details on the FACIT-F scale and its properties are presented in Appendix 3.

In both studies, the EORTC QLQ-C30 symptom scales were also assessed. The fatigue, dyspnea, and pain scale scores results are included in the present report and changes in the scores from baseline to week 26 were secondary end points. Scores range from 0 to 100 and higher symptom scale scores indicate a higher level of symptomology. Estimates of the MID for the scale scores in patients with cancer are 10 points to 20 points for moderate changes and greater than 20 points for large change. Estimates of MIDs for the scale scores were not identified in patients with PNH. More details on the EORTC QLQ-C30 and its properties can be found in Appendix 3.

In both studies, the presence or absence of the following signs and symptoms of PNH were recorded at each study visit: fatigue, chest pain, abdominal pain, dyspnea, dysphagia, erectile dysfunction, and red or dark urine or hemoglobinuria. Shifts in these signs and symptoms during the primary evaluation period were reported as secondary end points.

Breakthrough Hemolysis Events

If a suspected breakthrough hemolysis event occurred, LDH and other central laboratory assessments were to occur at an unscheduled visit (if the event did not occur at a scheduled visit). Breakthrough hemolysis was defined as at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, dyspnea, anemia [hemoglobin < 10 g/dL], MAVE, dysphagia, or erectile dysfunction) in the presence of LDH 2 × ULN or greater following prior reduction of LDH to less than 1.5 × ULN. Free C5 of 0.5 mcg/mL or greater was defined as suboptimal C5 inhibition and concomitant infections and complement-amplifying conditions were also assessed to further characterize breakthrough hemolysis events. Proportion of patients with breakthrough hemolysis during the primary evaluation period was a key secondary end point in both studies.

Intravascular Hemolysis

Serum LDH was assessed at screening and at each study visit in both studies (weekly until day 29, followed by visits every 2 weeks). Samples were collected before each study drug administration and sent for testing at a central laboratory. Samples with a serum potassium value of 6 mmol/L or greater and an LDH value of 2 × ULN or greater were considered to have undergone ex vivo hemolysis and were excluded from efficacy analyses, although such LDH values could be used for determining breakthrough hemolysis. LDH normalization was defined as an LDH level less than or equal to the ULN (246 U/L). Proportion of patients with LDH normalization from day 29 through day 183 was a coprimary end point in Study 301, percent change in LDH from baseline to day 183 was a key secondary end point in Study 301 and the primary end point in Study 302, and proportion of patients with LDH normalization at day 183 was a secondary end point in Study 302.

The relationship between the LDH threshold of 1.5 × ULN and the key PNH clinical outcomes of mortality and thromboembolism has been described in multiple publications reporting on patients in a national South Korean PNH registry who had not received eculizumab. Details on the evidence supporting 1.5 × ULN as a clinically meaningful threshold for LDH are presented in Appendix 3.

Hemoglobin Stabilization

Hemoglobin was assessed at screening and at each study visit in both studies. As with serum LDH, samples were collected before each study drug administration and sent for testing at a central laboratory. Proportion of patients with hemoglobin stabilization was a key secondary end point in both studies with stabilization defined as avoidance of a 2 g/dL or greater decrease in hemoglobin level in the absence of transfusion from baseline through day 183.

Adverse Events

AEs were assessed through continuous AE monitoring; laboratory results for hematology, blood chemistry, coagulation, and urinalysis measurements at each study visit; vital signs measurements at each study visit; and abbreviated physical examination and electrocardiograms on days 1, 71, and 183.

Immunogenicity

Samples for assessing presence and titre of ADAs were collected before dosing on days 1, 71, 127, and 183 in both studies.

Primary End Points of the Studies

The coprimary end points of Study 301 were percentage of patients with transfusion avoidance throughout the primary evaluation period (assessed by the difference between groups) and proportion of patients with LDH normalization from day 29 through day 183 (assessed by the OR). Ravulizumab had to meet noninferiority for both end points to meet the primary objective of the study. The primary end point of Study 302 was percent change in LDH from baseline to day 183 and, similarly, ravulizumab was tested for noninferiority to eculizumab.

For transfusion avoidance, the lower bound of the 95% CI for the OR of ravulizumab versus eculizumab had to be greater than 0.39 to meet the noninferiority margin. For LDH normalization, the lower bound of the 95% CI had to be higher than –20% for the difference in percentage of patients with LDH normalization in the ravulizumab group versus the eculizumab group to meet the noninferiority margin. For percent change in LDH, the upper bound of the 95% CI had to be lower than 15% for the mean percent change in LDH for ravulizumab versus eculizumab.

Statistical Testing

If the coprimary and primary end points in Study 301 and Study 302, respectively, were met, then the key secondary end points were also tested for noninferiority of ravulizumab using a closed testing procedure. If all key secondary end points demonstrated noninferiority, a hierarchy of end points was to be tested for superiority using a closed testing procedure and a significance level of 0.05. For the noninferiority and superiority testing hierarchies, please refer to Table 6. In each hierarchy, the next end point could only be tested if statistical significance was met in the previous end point.

Sample Size Calculations

In Study 301, the required sample size was based on the calculations for the coprimary end point requiring more patients. For the LDH normalization and transfusion avoidance end points, it was estimated that at least 142 patients and 193 patients would be required, respectively, to provide 80% power to demonstrate noninferiority. The estimated required sample size for Study 301 was 214, based on the transfusion avoidance end point and an assumed 10% dropout rate.

In Study 302, it was estimated that 192 patients would be required to provide 90% power to demonstrate noninferiority in percent change in LDH, assuming a SD of 30% in both treatment groups (based on the TRIUMPH study²¹), and a dropout rate of 10%.

There were no power calculations reported for key secondary or secondary end points in either study.

Noninferiority Margins

For all noninferiority tests, the lower bound of the 95% CI of the estimate of the difference for ravulizumab versus eculizumab had to be higher than the specified noninferiority margin (or the upper bound had to be lower than the margin). As presented in Table 9, most of the noninferiority margins were based on a 50% or less loss of benefit from eculizumab versus placebo in the TRIUMPH study²⁰ or a 50% loss of benefit based on patients treated with eculizumab versus untreated patients (for the Study 301 margins) or patients who discontinued eculizumab (for the Study 302 margins) in the sponsor’s global PNH registry. Some margins were slightly more conservative compared to a margin determined by a 50% or less loss of benefit. For percent change in LDH, the margins were based on a 25% or less loss of benefit in Study 301 and a 11% or less loss of benefit in Study 302. Although it was acknowledged that more conservative noninferiority margins could have been selected, it was determined that the required sample size for more conservative margins would not have been feasible give the rarity of PNH. In particular, it would have been difficult to enrol sufficient numbers of treatment-naive patients for Study 301.

Table 9

Noninferiority Margins.

Subgroup Analyses

There was no pre-specified statistical testing for differences in efficacy by subgroups of patients. The following subgroups were analyzed for primary, coprimary, and key secondary end points in both studies: randomization stratification variables (transfusion history for both studies and screening LDH level for Study 301), sex, race, region, and age. The only relevant subgroups in the studies, according to the CADTH systematic review protocol, were those categorized according to screening LDH level. In Study 301, the categories were screening LDH level of 1.5 to less than 3 × ULN and 3 × ULN or greater. There were no subgroups in Study 302 categorized by LDH level.

Data Imputation Methods

Data imputation methods for each outcome were the same across both studies. For percent change in LDH, proportion of patients with LDH normalization, and FACIT-F score there was no imputation for a patient at a particular visit. For proportion of patients with transfusion avoidance, breakthrough hemolysis, or hemoglobin stabilization, patients who discontinued the study due to lack of efficacy during the primary evaluation period were considered nonresponders. Patients who discontinued the study for other reasons were included, but assessments following discontinuation were not included.

For the EORTC QLQ-C30 subscale scores and the FACIT-F subscale score, scores were computed if more than 50% of the items were completed, in accordance with their scoring guidelines. There was no imputation for missing subscale scores.

Sensitivity Analyses

In addition to PP analyses, additional sensitivity analyses were conducted for the coprimary and primary end points. In Study 301, both coprimary end points were analyzed using a finer categorization of transfusion history (0, 1 to 4, > 4 to 14, or > 14 units), as well as no adjustment factors. Transfusion avoidance was also analyzed with response being defined as not actually receiving a transfusion as opposed to not meeting the transfusion criteria. LDH normalization was analyzed with transfusion history as a continuous variable as well as with a weighted GEE to account for drop-outs under the missing at random assumption. As well, the proportion of patients with a median LDH from day 29 through day 183 of below ULN was compared between treatment groups. In Study 301, the primary end point was also analyzed without adjustment for transfusion history and baseline LDH.

Statistical Models for Key Secondary End Points

A summary of the statistical models used for the primary and key secondary end points is presented in Table 10. Percent change in LDH in Study 301 and percent change in FACIT-F score in both studies were analyzed using the same methods as for the primary end point in Study 302 (with both randomization strata in Study 301 used in place of transfusion history as fixed, categorical effects). In Study 301, breakthrough hemolysis and hemoglobin stabilization were analyzed using the same methods as for transfusion avoidance. In Study 302, transfusion avoidance, breakthrough hemolysis, and hemoglobin stabilization were analyzed using the methods as for transfusion avoidance in Study 301, except that the Study 302 randomization stratum (transfusion history) replaced the Study 301 randomization strata as a fixed, categorical effect.

LDH normalization in Study 302, a secondary end point, was analyzed using the same model and adjustment factors as for LDH normalization in Study 301, except that LDH normalization from baseline to day 183 was considered (rather than starting at day 29).

In both studies, the EORTC QLQ-C30 subscale scores (secondary end points) were reported as summary statistics (mean, median, and change from baseline) at each assessment. Shifts from baseline in clinical manifestations of PNH in both studies, also secondary end points, were summarized for each study visit.

Table 10

Statistical Analysis of Efficacy End Points.

Analysis Populations

In both studies, the primary efficacy analyses were conducted in the FAS and patients were analyzed according to the treatment group to which they were randomized. In Study 301, the FAS included all patients who received at least 1 dose of study drug and had at least 1 efficacy assessment following the first dose of study drug. In Study 302, the FAS included all patients who received at least 1 dose of study drug.

The PP set included all patients in the FAS who missed no doses of ravulizumab or no more than 1 dose of eculizumab during the primary evaluation period, never received the wrong study treatment, followed the protocol-specified transfusion criteria, and met key study selection criteria.

The safety set in both studies was used for safety analyses and included all patients who received at least 1 dose of study drug. Patients were analyzed according to the study treatment they received.

Survival

Survival was not assessed in Study 301 and Study 302 as an efficacy outcome.

Thrombotic Events

Treatment-emergent MAVEs were assessed in Study 301 and Study 302. In Study 301, 2 patients in the ravulizumab group each experienced 1 event of deep vein thrombosis and 1 patient in the eculizumab group experienced a mesenteric venous thrombosis event. There were no MAVEs in Study 302.

Health-Related Quality of Life

The results for the EORTC QLQ-C30 global health status score are presented in Table 14. Change in the global health status score from baseline to week 26 was a secondary end point and not part of the closed testing procedure in both studies. In Study 301, patients in the ravulizumab and eculizumab groups had a change in global health status score of 13.17 (SD = 21.44) and 12.85 (SD = 21.83), respectively. In Study 302, baseline and week 26 scores were similar to each other within each group, with a change in global health status score of 1.15 (SD = 16.51) in the ravulizumab group and –1.93 (SD = 15.34) in the eculizumab group. Increase in global health status score corresponds to improvement in health status.

Table 14

EORTC QLQ-C30 Global Health Status Score.

Transfusions

The results for transfusion avoidance and transfusions are presented in Table 15. Transfusion avoidance was a coprimary end point in Study 301 and a key secondary end point in Study 302 that was tested for noninferiority in both studies according to the closed testing procedure. According to the protocol-specified transfusion criteria, the mean difference in percentage of patients achieving transfusion avoidance in the ravulizumab versus the eculizumab group was 6.8% (95% CI, –4.66% to 18.14%) in Study 301 and 5.5% (95% CI, –4.27% to 15.68%) in Study 302. Noninferiority was met in both studies as the lower bounds of the 95% CIs were higher than –20%. The results for the PP analyses, which excluded patients who met the transfusion criteria but did not receive a transfusion during the primary evaluation period, were consistent with the FAS results. The results from other sensitivity analyses using a different categorization of transfusion history, defining transfusion avoidance as achieved only by patients who did not receive a transfusion, and removing adjustment for randomization factors were also consistent with the FAS results.

In the subgroups categorized by screening LDH level in Study 301, mean difference in percentage of patients achieving transfusion avoidance in the ravulizumab versus the eculizumab group was 20.1% (95% CI, –13.37% to 51.04%) in patients with LDH of 1.5 to 3 × ULN and 5.3% (95% CI, –8.28% to 18.52%) in patients with LDH of 3 × ULN and greater.

In patients who received at least 1 transfusion, the mean number of transfusions in the ravulizumab and eculizumab groups was 3.3 (SD = 4.15) and 3.6 (SD = 3.06) in Study 301 and 2.7 (SD = 2.75) and 2.0 (SD = 1.29) in Study 302. The mean number of units transfused in the ravulizumab and eculizumab groups was 4.8 (SD = 5.06) and 5.6 (SD = 5.93) in Study 301 and 4.3 (SD = 4.76) and 3.4 (SD = 3.01) in Study 302.

Table 15

Transfusions and Transfusion Avoidance.

Symptoms of PNH

The results for the FACIT-F score are presented in Table 16. The change in FACIT-F total score was a key secondary end point and was tested for noninferiority in accordance with the closed testing procedure in both studies. The mean difference in change from baseline to week 26 in FACIT-F total score in the ravulizumab versus the eculizumab group was 0.67 (95% CI, –1.21 to 2.55) in Study 301 and 1.47 (95% CI, –0.21 to 3.15) in Study 302. Noninferiority was met in both studies as the lower bounds of the 95% CIs were higher than –5 and –3 in Study 301 and Study 302, respectively. The results for the PP analyses were consistent with the FAS results.

Table 16

Fatigue Measured by FACIT-F Total Score.

The results for the EORTC QLQ-C30 symptom subscale scores are presented in Table 17. Changes from baseline to week 26 in the subscale scores were secondary end points and were not part of the closed testing procedure in both studies. Decreases in symptom subscale scores correspond to improvement in symptoms. Changes in the treatment-naive patients of Study 301 were consistently greater in magnitude than changes in patients in Study 302. In Study 301, change for the ravulizumab and eculizumab groups, respectively, was –20.2 (SD = 24.51) and –18.6 (SD = 24.49) for the fatigue scale, –11.3 (SD = 23.43) and –7.6 (SD = 23.41) for the pain scale, and –14.1 (SD = 24.78) and –17.1 (SD = 25.99) for the dyspnea scale. In Study 302, change for the ravulizumab and eculizumab groups, respectively, was –4.97 (SD = 17.260) and –0.71 (SD = 15.271) for the fatigue scale, 0.87 (SD = 12.722) and 1.42 (SD = 19.959) for the pain scale, and –1.74 (SD = 18.950) and –1.06 (SD = 20.417) for the dyspnea scale.

Table 17

EORTC QLQ-C30 Symptom Subscale Scores.

The results for shifts in signs and symptoms of PNH recorded at baseline and at day 183 are presented in Table 18. In both studies, shifts during the primary evaluation period were secondary end points. For each of the signs and symptoms in the treatment-naive patients of Study 301, most of the patients with the sign or symptom at baseline had a shift such that they did not have the clinical manifestation at day 183. Shifts in signs and symptoms absent at baseline in Study 301 to being present at day 183 were reported in 5.0% of patients or less in each treatment group. For each of the signs and symptoms in the previously treated patients of Study 302, most patients did not report a shift from baseline to day 183. For fatigue, 20.8% of patients in the ravulizumab group reported no fatigue at baseline with a shift to fatigue being present at day 183 compared with 8.4% in the eculizumab group. All other clinical manifestations had 10.5% of patients or less with a shift from absent to present or from present to absent in each treatment group.

Table 18

Shifts in Sign and Symptoms of PNH.

Breakthrough Hemolysis Events

The results for breakthrough hemolysis are presented in Table 19. The proportion of patients with breakthrough hemolysis during the primary evaluation period was a key secondary end point in both studies. It was tested for noninferiority in both studies and for superiority in Study 301 in accordance with the closed testing procedure. The mean difference in the percentage of patients with breakthrough hemolysis in the ravulizumab versus the eculizumab group was –6.7% (95% CI, –14.21% to 0.18%) in Study 301 and –5.1% (95% CI, –18.99% to 8.89%) in Study 302. Noninferiority was met in both studies as the upper bounds of the 95% CIs were lower than 20%. The results for the PP analyses were consistent with the FAS results. Breakthrough hemolysis was the first outcome in the Study 301 testing hierarchy for superiority. The significance level was not met for superiority and no further superiority testing was performed.

A breakdown of events by whether they were concurrent with suboptimal C5 inhibition (free C5 ≥ 0.5 mcg/mL) and/or a complement-amplifying condition (i.e., infection) is provided in Table 19. Patients receiving ravulizumab in both studies did not experience a breakthrough hemolysis event concurrent with suboptimal C5 inhibition, while 7 and 4 patients receiving eculizumab in Study 301 and Study 302, respectively, did experience such events (with 2 of those patients in Study 301 and 1 of those patients in Study 301 having a concomitant infection).

Table 19

Breakthrough Hemolysis.

Complications of PNH

Complications of PNH were not assessed in Study 301 and Study 302 as an efficacy outcome.

Intravascular Hemolysis

The results for LDH normalization and mean LDH level are presented in Table 20. LDH normalization was a coprimary end point in Study 301 and a secondary end point in Study 302. The OR for the proportion of patients achieving LDH normalization from day 29 to 183 in Study 301 was 1.187 (95% CI, 0.796 to 1.769) for ravulizumab versus eculizumab. Noninferiority was met as the lower bound of the 95% CI was greater than 0.39. The results for the PP analysis and the other sensitivity analyses using different categorization of transfusion history, removing adjustment for randomization factors, and weighted GEE to account for drop-outs under the missing at random assumption were consistent with the main results. In patients with a screening LDH of 1.5 to 3 × ULN (N = 34), the OR was 1.71 (95% CI, 0.56 to 5.23) and in patients with a screening LDH of ≥ 3 × ULN (N = 212), the OR was 1.13 (95% CI, 0.73 to 1.73). In Study 302, the OR for the proportion of patients achieving LDH normalization from baseline to day 183 was 0.801 (95% CI, 0.500 to 1.282).

Mean percent change in LDH level from baseline to day 183 was the primary end point in Study 302 and a key secondary end point in Study 301. It was tested for noninferiority in both studies and for superiority in Study 302 in accordance with the closed testing procedure. In Study 302, the least squares mean difference in percent change in LDH level was –9.21% (95% CI, –18.84% to 0.42%) for ravulizumab versus eculizumab. Noninferiority was met as the upper bound of the 95% CI was lower than 15%. Percent change in LDH was the first outcome in the Study 302 testing hierarchy for superiority. The significance level was not met for superiority and no further testing was performed. In Study 301, the least squares mean difference in percent change in LDH level was –0.83% (95% CI, –5.21% to 3.56%) for ravulizumab versus eculizumab. Noninferiority was met as the upper bound of the 95% CI was lower than 20%.

Table 20

LDH Normalization and LDH Level.

Hemoglobin Stabilization

The results for hemoglobin stabilization are presented in Table 21. Hemoglobin stabilization throughout the primary evaluation period was a key secondary end point for both studies and was tested for noninferiority in both studies according to the closed testing procedure. The mean difference in percentage of patients with hemoglobin stabilization in the ravulizumab versus the eculizumab group was 2.9% (95% CI, –8.80% to 14.64%) in Study 301 and 1.4% (95% CI, –10.41% to 13.31%) in Study 302. Noninferiority was met in both studies as the lower bounds of the 95% CIs were higher than –20%. The results for the PP analyses, which excluded patients who met the transfusion criteria but did not receive a transfusion during the primary evaluation period, were consistent with the FAS results.

Table 21

Hemoglobin Stabilization.

Health Care Resource Utilization

Health care resource utilization was not assessed in Study 301 and Study 302 as an efficacy outcome.

Adverse Events

Most patients (86.8% to 88.0%) in both treatment groups in both studies reported at least 1 AE. The most common AE was headache, which was reported in 17.3% to 36.0% of patients in each treatment group. In Study 302, headaches were reported by 26.8% of the ravulizumab group and 17.3% of the eculizumab group. According to the clinical expert consulted by CADTH, it is possible that the transition from pre-study eculizumab to ravulizumab accounted for this difference due to the more profound C5 blockade with ravulizumab. Headaches were numerically more common in both treatment groups in the treatment-naive patients of Study 301 compared with the ravulizumab group of Study 302. Other AEs that were reported in at least 10% of any treatment group were upper respiratory tract infection, nasopharyngitis, pyrexia, and cough. There were no notable imbalances in AEs.

Serious AEs

SAEs were reported in 4.1% to 8.8% of each treatment group in both studies. The most common SAEs were hemolysis and pyrexia, which occurred in 3.1% or less of each treatment group.

Withdrawals Due to AEs

There were no withdrawals due to AEs in either study.

Mortality

One patient in the eculizumab group in Study 301 died due to lung adenocarcinoma during the extension phase of the study.

Notable Harms

Serious infections were reported in 1.0% to 3.3% of each treatment group in both studies. Each specific SAE was reported in 1 patient. Infusion reactions were reported in 3.1% to 8.8% of patients across each treatment group in both studies. In Study 301, the AEs leading to infusion interruption were 1 infusion-related reaction (lower back pain) and 1 patient with muscle spasms during each infusion in the ravulizumab group and 1 patient with recurrent headache in the eculizumab group. In Study 302, 1 patient had back pain (3 occurrences) and 1 patient had flank pain, both in the eculizumab group.

In Study 301, ADAs were present in 12 patients (9.6%) in the ravulizumab group and in 6 patients (5.0%) in the eculizumab group at baseline and there was 1 treatment-emergent ADA-positive sample in each treatment group. In Study 302, ADAs were present in 4 patients (4.1%) in the ravulizumab group at baseline and there was 1 treatment-emergent ADA-positive sample in the eculizumab group. All ADA titres in the positive samples were considered to be low.

Table 22

Summary of Harms.

Internal Validity

Study 301 and Study 302 had appropriate randomization and allocation methods and there were no notable imbalances in baseline characteristics between the treatment groups. Study discontinuations were 3% or less in each treatment group and treatment adherence was 100% for all but 1 patient across both studies. Therefore, there are no concerns of bias due to study drop-outs or imbalanced treatment exposure.

The statistical analysis methods used for the efficacy end points were appropriate. The closed testing procedure in both studies was appropriate for controlling the type I error rate and ensuring an accurate interpretation of statistical significance for tests of superiority.

Justification for the chosen noninferiority margins was provided based on prior RCT and registry data; however, the margin for all end points aside from percent change in LDH was based on a 50% or less loss of benefit or similar loss of benefit from eculizumab. The rationale for not selecting more conservative noninferiority margins was that the required sample size would not have been feasible given the rarity of the disease. The clinical expert consulted by CADTH for this review noted that a 50% loss of benefit would not be clinically acceptable, but also recognized that the rarity of the disease does have implications for clinical trial recruitment. The margin for percent change in LDH was based on a 25% or less loss of benefit in Study 301 and a 11% or less loss of benefit in Study 302 and the clinical expert consulted by CADTH found these margins more clinically acceptable. Concerns about the relatively generous noninferiority margins are alleviated by the fact that all primary and key secondary end points met the noninferiority margin in both the primary and PP analyses and that in all cases a more conservative margin would have been met.

Although the subgroup analyses in the studies were preplanned, there were no sample size considerations for them, no control for type I error rate, and no statistical testing for treatment-by-subgroup interaction. Therefore, conclusions cannot be drawn regarding analyses by subgroup.

The open-label nature of the studies means that outcomes relying on subjective reporting, such as the EORTC QLQ-C30 and the FACIT-F could have been biased, with potential for bias in favour of ravulizumab. Determination of the presence of PNH-related signs or symptoms was not considered by the clinical expert to be prone to bias from lack of blinding. Therefore, end points based on laboratory values and/or presence of PNH-related signs or symptoms (i.e., all other primary and key secondary end points) had a low likelihood of bias from lack of blinding. However, reporting of AEs may have been susceptible to bias from lack of blinding.

While the EORTC QLQ-C30 and the FACIT-F have been used extensively in other disease areas, their reliability, validity, and responsiveness have yet to be characterized in patients with PNH. Although estimates for MIDs for these instruments were not identified in this patient population, changes in the relevant scores were not tested for superiority.

External Validity

The patient populations in the studies were overall representative of patients seen in Canadian clinical practice, according to the clinical expert consulted by CADTH. However, there were some exclusions of small subpopulations of patients due to the study selection criteria. The clinical expert noted that patients with frank bone marrow failure would likely have been excluded due to the criteria on platelet and absolute neutrophil count, though these patients in clinical practice can still receive treatment. The criteria for Study 302 were chosen in such a way that patients requiring a higher dose or more frequent dosing of eculizumab beyond the product monograph-recommended dosage would have been excluded, according to the clinical expert. Also, patients who were pregnant or breastfeeding were excluded.

As outlined by the clinical expert in the clinical input section earlier in the present report, approximately 20% of patients require higher doses of eculizumab or more frequent doses than recommended in the Health Canada–approved product monograph for the PNH indication (900 mg maintenance dose every 2 weeks) to properly control symptoms and signs associated with pharmacokinetic breakthrough. Therefore, a similar portion of patients in Study 301 would have been expected to experience pharmacokinetic breakthrough in the eculizumab group since the studies did not allow for deviation from the labelled dosage of eculizumab. This is contrary to Canadian clinical practice, in which dosing would be adjusted to address pharmacokinetic breakthrough. For Study 301, the lack of permitted dosage adjustments may have biased the efficacy results in favour of ravulizumab relative to how eculizumab is dosed in clinical practice.

Finally, the efficacy results do not address important outcomes such as survival and complications of PNH other than thrombotic events (not assessed in either study) and MAVEs (no statistical testing).

Methods

Presented are the results from the extension periods of both Study 301 and Study 302. Following the 26-week randomized treatment period, patients had the option to continue into the extension period where all patients received ravulizumab for an additional 26 weeks. Patients in the ravulizumab treatment arm continued receiving ravulizumab maintenance dosing every 8 weeks, while patients in the eculizumab arm were transitioned to receive a ravulizumab loading dose followed by ravulizumab maintenance dosing every 8 weeks for an additional 26 weeks.

Populations

All patients who entered the extension period of both studies were originally enrolled in the randomized treatment period of Study 301 and Study 302. As such, the inclusion and exclusion criteria are consistent with what is reported in earlier sections. Of the 246 patients who received ravulizumab or eculizumab in Study 301, 243 (98.8%) patients entered the extension period, and of the 192 patients who received ravulizumab or eculizumab in Study 302, 191 (99.4%) patients entered the extension period. Therefore, baseline characteristics are expected to be similar as reported in the systematic review section.

Interventions

All patients either maintained their treatment with ravulizumab or were transitioned from eculizumab to ravulizumab. Use of the study drug was consistent with the primary evaluation period.

Outcomes

The primary and key secondary end points assessed in the primary evaluation period of both studies were also assessed during the extension period of both studies. Briefly, they were transfusion avoidance, FACIT-F scores, breakthrough hemolysis events, LDH normalization and change from baseline, hemoglobin stabilization, as well as safety outcomes.

Statistical Analysis

There was no comparator arm in the extension period as all patients either maintained or were transitioned to treatment with ravulizumab. As such, there was no formal statistical testing conducted on the results. All results presented are for the FAS of patients who entered the extension period and descriptive statistics were used to summarize the results rather than the statistical models used in the primary evaluation period. The initial 26-week primary evaluation period will be referred to as Period 1 and the following 26-week extension period will be referred to as Period 2.

Patient Disposition

In Study 301, 246 total patients were treated with either ravulizumab or eculizumab in the 26-week primary evaluation period. Of these, 99.2% of the patients in the ravulizumab group and 98.3% of the patients in the eculizumab group entered the extension period with 2.4% and 4.1%, respectively, discontinuing treatment at some point during the extension period. Two patients from the eculizumab to ravulizumab group discontinued due to an AE. Other reasons for discontinuation included pregnancy, physician decision, as well as 1 patient who died while on treatment.

In Study 302, 195 total patients were treated with either ravulizumab or eculizumab in the 26-week primary evaluation period. Of these, 99.0% of the patients in the ravulizumab group and 96.9% of the eculizumab group entered the extension period, with 1% of patients in each group discontinuing at some point during the extension period due to either physician decision or withdrawal by patient. Table 23 below reports patient disposition results for both Study 301 and Study 302.

Table 23

Patient Disposition in Extension Studies.

Exposure to Study Treatments

As Study 301 and Study 302 continued to monitor patients beyond the 52-week time point presented in this review, mean treatment duration is greater than noted during the primary evaluation period with a mean duration of 290.7 days for both groups in Study 301. In Study 302 there was shorter follow-up available with a mean duration of 216.5 days (SD = 31.06) in the ravulizumab to ravulizumab group and 216.0 days (SD = 27.83) in the eculizumab to ravulizumab group. A similar difference is seen between Study 301 and Study 302 with regards to number of infusions per patient. Infusion interruptions were more commonly seen in Study 301 with a total of 17 across both groups, of these 4 were reported to be due to AEs. Treatment adherence was consistent across studies and treatment groups. Full treatment exposure results are shown in Table 24.

Table 24

Treatment Exposure in the Extension Period.

Survival

Survival was not assessed in the Study 301 and Study 302 extensions as an efficacy outcome.

Thrombotic Events

In Study 301, there were 2 MAVEs reported during the extension period up to the 52-week time point. One event was an arterial embolism and the second was a jugular vein thrombosis. In Study 302, there were 2 MAVEs reported during the extension period, 1 of thrombophlebitis and 1 cerebral infarction.

Results for MAVEs occurring during the 2 years before ravulizumab initiation and up to 2 years following ravulizumab initiation in Study 301 are reported in an abstract presentation from Latour et al.²⁴ The results are not included here due to the post hoc nature of the analysis and the immature follow-up at the 2-year extension time point.

Health-Related Quality of Life

HRQoL as measured by EORTC QLQ-C30 was not reported in the Study 301 and Study 302 extensions.

Transfusions

Table 25 presents the transfusion avoidance results for Study 301 and Study 302. The number of patients achieving transfusion avoidance in the extension period remained consistent with the results from the primary evaluation period in both studies. In Study 301, 90.2% of patients in the ravulizumab to ravulizumab group and 87.3% of patients in the eculizumab to ravulizumab group maintained response, while 62.5% and 72.5% of patients in each group who did not respond in the primary evaluation period maintained no response, with regards to transfusion avoidance, during the extension period.

Table 25

Transfusion Avoidance in the Extension Period.

Symptoms of PNH

Fatigue as measured by FACIT-F score is shown below in Figure 2 for Study 301, and Figure 3 for Study 302. FACIT-F scores observed at the end of the primary evaluation period appeared to be maintained throughout the extension period in both Study 301 and Study 302.

Figure 2

Fatigue Measured by FACIT-Fatigue (Study 301 Extension).

Figure 3

Fatigue Measured by FACIT-Fatigue (Study 302 Extension).

Breakthrough Hemolysis

Table 26 presents the breakthrough hemolysis results for Study 301 and Study 302. The results for breakthrough hemolysis in the extension period were similar to the results from the primary evaluation period in both studies. The group that transitioned from eculizumab to ravulizumab saw a numeric reduction in proportion of patients with breakthrough hemolysis from 10.9% in Period 1 to 1.6% in Period 2 in Study 301 and from 4.2% in Period 1 to 1.1% in Period 2 in Study 302. The group that maintained treatment with ravulizumab into Period 2 saw similar proportions of patients with breakthrough hemolysis from 4.0% in Period 1 and 3.2% In Period 2 in Study 301, and a slight numeric increase from 0% in Period 1 to 3.1% in Period 2 in Study 302. In Study 301, 99.2% of patients in both the ravulizumab to ravulizumab group and eculizumab to ravulizumab group who did not experience a breakthrough hemolysis event in Period 1, maintained no breakthrough hemolysis events in Period 2. Similarly, in Study 302, 96.9% of patients in the ravulizumab to ravulizumab group and 100% of patients in the eculizumab to ravulizumab arm, maintained a no breakthrough hemolysis event response.

Table 26

Breakthrough Hemolysis in the Extension Period.

Intravascular Hemolysis

Table 27 presents the LDH normalization results for Study 301. The results for LDH normalization in the extension period remained consistent with the results from the primary evaluation period. In Study 301, 75% of patients in the ravulizumab to ravulizumab group and 72% of patients in the eculizumab to ravulizumab group who had achieved LDH normalization in Period 1 maintained normalization in Period 2. Conversely, 85% of patients in the ravulizumab to ravulizumab group and 80.6% of patients in the eculizumab to ravulizumab group who did not achieve LDH normalization in Period 1 maintained a lack of response throughout Period 2. Figure 4 below shows the LDH levels visually throughout Study 301, indicating that mean LDH levels at the end of the primary evaluation period were maintained in both groups throughout the extension period.

Table 27

LDH Normalization in the Extension Period.

Figure 4

LDH Levels Over Time (Study 301 Extension).

Figure 5 shows the mean LDH levels in Study 302 for both groups throughout the primary evaluation period and extension period. LDH levels in the extension period remained consistent with the results found in the primary evaluation period.

Figure 5

LDH Levels Over Time (Study 302 Extension).

Hemoglobin Stabilization

Table 28 presents the hemoglobin stabilization results for Study 301 and Study 302. The results for hemoglobin stabilization in the extension period remained consistent with the results from the primary evaluation period in both studies. In Study 301, 89.4% of patients in the ravulizumab to ravulizumab group and 85.7% of patients in the eculizumab to ravulizumab group who achieved hemoglobin stabilization in Period 1 maintained their response in Period 2. Similarly, in Study 302, 81.2% of patients in the ravulizumab to ravulizumab group and 81.1% of patients in the eculizumab to ravulizumab arm who achieved hemoglobin stabilization in Period 1 maintained their response in Period 2.

Table 28

Hemoglobin Stabilization in the Extension Period.

Harms

Presented in Table 29 is the summary of harms from the extension period up to 52 weeks for Study 301 and Study 302. No additional safety signals were identified when compared to the results presented for the 26-week primary evaluation period. In Period 2 of Study 301, 63.7% of patients in the ravulizumab to ravulizumab group and 74.8% of patients in the eculizumab to ravulizumab group experienced 1 or more AEs. In Period 2 of Study 302, 79.2% of patients in the ravulizumab to ravulizumab group and 74.7% of patients in the eculizumab to ravulizumab group experienced 1 or more AEs. Similar to what was seen in the primary evaluation period, headaches and nasopharyngitis were the most common AEs reported; however, the frequency was consistently lower in Period 2 compared to the primary evaluation period across all groups. Patients experiencing 1 or more SAEs in both the Study 301 and Study 302 extension periods was more consistent with the results from the primary evaluation period. In Study 301, 7.3% of patients in the ravulizumab to ravulizumab group and 5.9% of patients in the eculizumab to ravulizumab group experienced an SAE, while in Study 302 the percentages of patients were 8.3% and 5.3%, respectively.

Of the notable harms identified, serious infections appeared to remain consistent with the results from the primary evaluation period while infusion reactions were slightly lower in Period 2 compared to the primary evaluation period. In Study 301, 1.6% of patients in the ravulizumab to ravulizumab group and 3.4% of patients in the eculizumab to ravulizumab group experienced a serious infection, while in Study 302 the percentages of patients were 3.1% and 2.1%, respectively. In Study 301, 4.0% of patients in the ravulizumab to ravulizumab group and 5.9% of patients in the eculizumab to ravulizumab group experienced an infusion reaction, while in Study 302 the percentages of patients were 2.1% and 1.1%, respectively.

The sponsor included safety data for 2 additional 6-month periods beyond the 52-week data cut-off; however, due to the immature follow-up for these time points, the results were not included in this review. From the sample size available, there were no additional noteworthy safety signals identified in these additional periods.

Table 29

Summary of Harms in the Extension Period.

Critical Appraisal

Methods

The substudy investigators composed a novel patient preference survey for the purposes of this study. To inform the creation of the survey, patient concept elicitation interviews were conducted in patients diagnosed with PNH who had previously received either ravulizumab or eculizumab, with preference for patients who were already participating in Study 302. The investigators finalized the questionnaire to include 11 questions in total; these include 1 overall preference question, 1 question evaluating preference according to 9 treatment characteristics, 1 question evaluating the most important characteristic impacting preference, 4 questions evaluating specific aspects of treatment with ravulizumab, and 4 questions evaluating specific aspects of treatment with eculizumab. The final 4 questions for each treatment used an agreement scale with 0 indicating no agreement with a given statement and 4 indicating full agreement with a given statement. All other preference questions were asked with the available responses of prefer ravulizumab, prefer eculizumab, or no preference.

Patients were given the opportunity to enrol in the patient preference substudy if they had also entered the extension period of Study 302 and had received a minimum of 2 doses of ravulizumab during the extension period. There were no exclusion criteria and the patient preference survey was administered to each participant at a single point in time.

Statistical Analysis

A planned sample size of 95 was calculated to have at least 80% power to detect a 65% or greater observed proportion of patients preferring ravulizumab under a null hypothesis of 50%. A 2-sided exact binomial test allowing for type I error of 0.05 was used in these calculations. Tests for mean differences in response were conducted for the final 4 questions for each treatment, P values were calculated from a paired t-test as well as the Wilcoxon signed rank test. There was no statistical hierarchy established and the analysis was not adjusted for multiplicity.

Results

Of the 98 patients that enrolled in the substudy, 3 did not respond to the first question of the survey, leaving 95 evaluable patients. The characteristics of the surveyed patients are shown in Table 30. Patients were evenly represented from both treatment arms with 53% of patients having received ravulizumab in the primary evaluation period and 47% having received eculizumab. The mean number of days between the last randomized study treatment and administration of the survey was 306 (SD = 55).

Table 30

Patient Characteristics in Study 302 Substudy.

Table 31 shows the results from the overall preference question as well as questions relating to specific aspects of treatment. When asked for overall preference, 93% of patients indicated that they preferred ravulizumab. For the treatment aspect questions, patients responded with either no preference, ranging from 2% to 53%, or a preference for ravulizumab, ranging from 45% to 98%. Two questions resulted in 98% of patients preferring ravulizumab — the questions regarding frequency of infusions and being able to plan activities.

Table 31

Treatment Preference in Study 302 Substudy.

Figure 6 shows which factors patients indicated were the most important for deciding treatment preference. Most patients chose frequency of infusions as the most important factor with overall quality of life the second most common factor cited by patients in the substudy.

Figure 6

Patients’ Most Important Treatment Factor for Deciding Preference in Study 302 Substudy.

Shown in Table 32 are 4 questions that were asked of patients regarding their experience receiving ravulizumab and eculizumab therapy. Patients generally favoured ravulizumab with the greatest absolute difference between the mean response scores observed in the question regarding frequency of infusions.

Table 32

Ratings on Treatment-Related Factors in Study 302 Substudy.

Critical Appraisal

ClinicalTrials.gov

Produced by the US National Library of Medicine. Targeted search used to capture registered clinical trials.

[Search -- Studies with results | Ultomiris or ravulizumab]

EU Clinical Trials Register

European Union Clinical Trials Register, produced by the European Union. Targeted search used to capture registered clinical trials.

[Search terms -- Ultomiris or ravulizumab]

Description and Scoring

The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) is one of the most used patient-reported outcome measures in oncology clinical trials. It is a multidimensional, cancer-specific, self-administered measure of HRQoL.²⁶

The EORTC QLQ-C30 is composed of both multi-item scales and single-item measures. These include 5 functional scales (physical, role, cognitive, emotional, and social), 3 symptom scales (fatigue, pain, and nausea and vomiting), a global health status/HRQoL scale, and 6 single items assessing additional symptoms commonly reported by cancer patients (dyspnea, loss of appetite, insomnia, constipation and diarrhea) as well as perceived financial impact of the disease.²⁶

The EORTC QLQ-C30 uses a 1-week recall period to assess functional status and symptoms. Most questions have 4 response options (“not at all,” “a little,” “quite a bit,” “very much”), with scores on these items ranging from 1 to 4. For the 2 items that form the global HRQoL scale, the response format is a 7-point Likert-type scale with anchors at 1 = “very poor” and 7 = “excellent.” Raw scores for each scale are computed as the average of the items that contribute to a particular scale. Each raw scale score is converted to a standardized score that ranges from 0 to 100 using a linear transformation. A higher score on the functional scales represents better functioning, a higher score on the symptom scales represents a higher level of symptomatology, and a higher score on the global health status/HRQoL scale represents a higher HRQoL.³⁴

According to the EORTC QLQ-C30 scoring algorithm, if there are missing items for a scale (i.e., the participant did not provide a response), the score for the scale can still be computed if there are responses for at least half of the items. The values for missing items are interpolated with the average of the respondent-completed items.³⁴

Assessment of Validity and Reliability

In its initial development, the EORTC QLQ-C30 underwent an evaluation of its psychometric properties and demonstrated reliability and validity in cancer patients in an international field trial of 305 patients in 13 multicultural clinical research settings.²⁶ A revision of the EORTC QLQ-C30 was undertaken to improve low internal consistency, content validity for the role functional scale, and a conceptual difficulty (undue emphasis on physical function in the global HRQoL scale).³⁵ The original and new versions were applied in a total of 1,181 patients with cancer in Canada and the Netherlands. Internal consistency improved for the role functioning scale in the new version (Cronbach alpha ranging from 0.78 to 0.88 in the 2 country samples), and substitution of the new item for the previous version did not alter internal consistency (Cronbach alpha ranging from 0.81 to 0.92).³⁵

The EORTC QLQ-C30 (version 3.0) is the version currently in use. Version 3.0 differs from the previous version 2.0 in that the number of response options for the first 5 items of the questionnaire comprising the physical function scale was increased from 2 options (yes/no in version 2.0) to 4 options (not at all, a little, quite a bit, very much). Internal consistency, reliability, construct validity, criterion validity, and responsiveness of the EORTC QLQ-C30 version 3.0 was assessed in 622 patients with head and neck cancer from 12 countries. Version 3.0 was more reliable than previous versions.³⁶ Internal consistency of the multi-item scales was assessed using Cronbach alpha, with a value of 0.70 being considered adequate.³⁷ The internal consistency of the new physical function scale of the EORTC QLQ-C30 version 3.0 was 0.84 compared with 0.66 in version 1.0. The EORTC QLQ-C30 version 3.0 was able to discriminate between head and neck cancer patients who were disease-free, who were newly diagnosed, and who had recurrent disease. As well, differences were noted between patients with different stages of disease and according to Karnofsky performance status (KPS): the new scale had a stronger association with KPS. Furthermore, there was a strong correlation observed between all subscale scores on the EORTC QLQ-C30 version 3.0 and symptom/treatment toxicity scores. Responsiveness to change was assessed using the standardized response mean (SRM), with an SRM of 0.20 considered small, 0.50 considered medium, and 0.80 considered large. The changes in the scores of QLQ-C30 demonstrated a small to medium SRM in response to treatment over time with scores mostly changing between 5 and 10 points.³⁶

In the Georgakopoulos et al. (2013) study, the validity of the EORTC QLQ-C30 was assessed in 80 newly diagnosed patients with Hodgkin lymphoma and diffuse large B-cell lymphoma (DLBCL) undergoing chemotherapy (Adriamycin, Bleomycin, Vinblastine, Dacarbazine [ABVD] for Hodgkin lymphoma, and R-CHOP for DLBCL).²⁷ Data were collected from the clinical research section of the Biomedical Research Foundation of the Academy of Athens in patients who had completed their chemotherapy (4-8 ABVD cycles or 6-8 R-CHOP cycles). The QLQ-C30 and other questionnaires were administered for self-completion, and the researcher was present for any clarifications. A difference of more than 10 units was considered significant for the 0-100 scales. Reliability as measured by Cronbach alpha for the EORTC QLQ-C30 was 0.79 for global health status/QoL, ranged from 0.51 to 0.85 for functional scales, and 0.82 to 0.86 for symptom scales/items indicating acceptable internal consistency for most dimensions. However, in the 2 functional scales of the QLQ-C30 instrument (emotional and cognitive functioning) the threshold of 0.70 was not met (0.63 and 0.51), demonstrating concerns around the internal consistency and reliability for these domains. No statistically significant differences between patients with HL and those with DLBCL were recorded, with exception in the symptom scale of the QLQ-C30 “appetite loss,” where a statistically significantly higher score for patients with HL was observed.²⁷

Evidence of validity and reliability of the EORTC QLQ-C30 was not identified in the literature for patients with PNH.

Minimal Important Difference

A study by Osoba and colleagues, conducted in patients with breast cancer and small-cell lung cancer, estimated that a change in score on any scale of the EORTC QLQ-C30 of 10 points would be clinically significant. This estimate was based on an anchor-based approach to estimate the MID in which patients who reported “a little” change (for better or worse) on the subjective significance questionnaire had corresponding changes on a function or symptom scale of the EORTC QLQ-C30 of approximately 5 to 10 points. Participants who reported a “moderate” change had corresponding changes in the EORTC QLQ-C30 of about 10 to 20 points, and those who reported “very much” change had corresponding changes in the EORTC QLQ-C30 of more than 20 points, though this was conducted in a previous version of the measure.²⁸

A more recent study from 2019 aimed to describe the MID for interpreting the EORTC QLQ-C30 measure in patients with advanced breast cancer patients. This study used an anchor-based approach utilizing performance status and chosen selected AEs as the clinical anchoring variables. The authors found that MIDs for within-group changes ranged from 5 to 14 points for improvements and from -14 to -4 points for deterioration across the individual scales. For between- group differences, the MIDs ranged from 4 to 11 points for improvements and from -18 to -4 points for deterioration across the individual scales.³⁸

No studies reporting MID in PNH were identified.

Description and Scoring

The Functional Assessment of Chronic Illness Therapy – Fatigue (FACIT-Fatigue) scale, previously known as the Functional Assessment of Cancer Therapy – Fatigue (FACT-F), is used to assess patient fatigue and energy levels and has been adapted for use in a number of chronic diseases.³⁹ It is a 13-item questionnaire and scores questions on a 5-point Likert scale. The total 13-item scale ranges from zero (extreme fatigue) to 52 (no fatigue).²⁹

Assessment of Validity and Reliability

In the initial development of the FACIT-Fatigue scale the questionnaire showed good stability (r = 0.87) along with strong internal consistency (Cronbach alpha = 0.95).²⁹ When analyzed for test-retest reliability coefficient the measure again showed good stability (r = 0.84 to 0.90) and strong internal consistency (Cronbach alpha = 0.93 to 0.95).²⁹ A more recent study from 2007 confirmed these results in a population of 135 patients with psoriatic arthritis that showed the FACIT-Fatigue questionnaire to have strong internal validity (Cronbach alpha = 0.95) and test-retest reliability (intraclass correlation coefficient = 0.95).³⁰ In study of 3,492 participants, the FACIT-Fatigue questionnaire was able to successfully discriminate anemic cancer patients from the general population with high sensitivity and reasonable specificity.⁴⁰ In a study of iron deficient anemia patients, FACIT-Fatigue showed convergence (r = 0.74) as well as responsiveness when compared to other relevant measures such as the SF-36 vitality subscale.⁴¹

No studies assessing the validity and reliability of the FACIT-Fatigue questionnaire in PNH were identified.

Minimal Important Difference

There were no identified studies that reported an MID for the FACIT-Fatigue scale in PNH.

Description

The screening and diagnosis of PNH relies on flow cytometry to identify PNH clones, presentation of clinical symptoms, and supportive laboratory tests. These laboratory tests can include hemoglobin, RBC counts, LDH levels.³¹ LDH normalization and change from baseline LDH are used as a primary surrogate outcome in both the Study 301 and Study 302 presented in the current review. Specifically, the threshold of LDH ≥1.5 x ULN is used as a key indicator for disease control.

Evidence of Relationship to Clinical Outcomes

The relationship between the LDH threshold of ≥1.5 x ULN and the key PNH clinical outcomes of mortality and thromboembolism has been described in multiple publications reporting on a national South Korean PNH registry covering 41 years of data and 301 patients who had not received eculizumab.^32,33 The Jang et al. (2016) study reported that, when compared to age- and gender-matched controls in the general population, a diagnosis of PNH resulted in a 3.9-fold increased risk of mortality.³² One of the key risk factors identified by the authors was patients with LDH ≥1.5 x ULN, which resulted in a 4.8-fold increase in risk of mortality compared to age- and gender-matched controls in the general population (p <0.001). PNH patients below this threshold showed no difference in mortality compared to healthy age- and gender-matched controls.³² The most important risk factor for mortality identified by the Jang et al. (2016) study was thromboembolism, which was associated with a 14-fold increased risk of mortality (p <0.001).³² Another publication using the same South Korean PNH registry reported on risk factors for thromboembolism in PNH patients.³³ Multivariate analysis of the dataset found that PNH patients with LDH levels ≥1.5 x ULN were at significantly increased risk of thromboembolism compared to patients below this LDH threshold (OR = 7.0; p = 0.013).³³ Patients with elevated LDH levels combined with clinical symptoms were associated with an even greater risk of thromboembolism than either single risk factor alone.³³ The results from these registry-based publications, along with general acceptance from Canadian clinicians, further support the use of LDH ≥1.5 x ULN as a relevant outcome in PNH.³