Patient-reported improvements in paroxysmal nocturnal hemoglobinuria treated with iptacopan from 2 phase 3 studies

Paroxysmal nocturnal hemoglobinuria (PNH) is a rare disease characterized by hemolysis and thrombosis and often occurs in the context of bone marrow failure.^1,2 PNH signs and symptoms, which are predominately caused by chronic intravascular hemolysis, can lead to anemia, fatigue, dyspnea, hemoglobinuria, impaired renal function, abdominal pain, dysphagia, and erectile dysfunction.^3,4 Fatigue is often cited as the most common and debilitating symptom of PNH, with 80.9% of patients in the International PNH Registry reporting fatigue.^3,5 

The IV administered C5 inhibitors (C5is) eculizumab and ravulizumab have significantly improved outcomes for patients with PNH. Both demonstrate control of intravascular hemolysis, with hemoglobin (Hb) stabilization in ∼50% of patients, reduced risk of thrombosis, and improved long-term survival.^6-11 However, many C5i-treated patients may experience persistent anemia, fatigue, and ongoing disease burden, mainly due to emerging extravascular hemolysis and, in some patients, residual intravascular hemolysis.^8,12-15 

Iptacopan is a first-in-class, oral, selective complement factor B inhibitor recently approved for treating adults with PNH.^16,17 Iptacopan targets the complement system proximally via the alternative pathway.^17,18 The primary results of active comparator–controlled APPLY-PNH for C5i-experienced patients with persistent anemia and single-arm APPOINT-PNH for C5i-naive patients have been reported.¹⁸ These 2 phase 3 trials demonstrated the efficacy and safety of iptacopan, with clinically meaningful improvements in Hb levels, control over both intravascular and extravascular hemolysis, and a favorable safety profile.¹⁸ 

Fatigue can substantially affect health-related quality of life (HRQOL)³; therefore, the patient-perceived benefit of therapy and the effects of therapy on the symptoms and impact of fatigue is an important goal for the treatment of PNH. In APPLY-PNH and APPOINT-PNH, fatigue and HRQOL were measured with patient-reported outcomes (PROs) including the Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-Fatigue),¹⁹ European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30), and Patient Global Impression of Severity (PGIS).^18,19 Meaningful within-patient change (MWPC) thresholds of response for the FACIT-Fatigue²⁰ and 4 subscales of the EORTC QLQ-C30 (physical functioning, role functioning, fatigue, and dyspnea) were determined. These thresholds were used to assess the proportion of patients who achieved MWPC in patient-reported fatigue severity and HRQOL during the 168-day core treatment period. The association between FACIT-Fatigue scores and hematologic markers of disease control (ie, lactate dehydrogenase [LDH] and Hb) in both APPLY-PNH and APPOINT-PNH was also examined.

APPLY-PNH (ClinicalTrials.gov identifier: NCT04558918) was an open-label, randomized, multicenter, phase 3 trial investigating iptacopan monotherapy in patients with PNH and persistent anemia despite C5i therapy (eculizumab or ravulizumab). APPOINT-PNH (ClinicalTrials.gov identifier: NCT04820530) was an open-label, multicenter, phase 3 trial investigating iptacopan monotherapy in C5i-naive patients with PNH. Both trials included a 56-day (8-week) screening period, a 168-day (24-week) core treatment period, and a 168-day (24-week) extension period. Iptacopan was self-administered at 200 mg twice daily in both studies.¹⁸ In APPLY-PNH, C5is were administered IV.^18,21,22 

Details of the study design, inclusion and exclusion criteria, and primary analyses have been described previously.¹⁸ Briefly, both trials enrolled adult patients (age of ≥18 years) with PNH, mean Hb of <10 g/dL, and no laboratory evidence of bone marrow failure. In APPLY-PNH, patients were previously treated with a stable regimen of C5i therapy (eculizumab or ravulizumab) for ≥6 months before randomization. In APPOINT-PNH, patients were C5i naive and had LDH level of >1.5 × the upper limit of normal.¹⁸ 

The protocols were approved by the institutional review board or ethics committee at each participating study site.

In both APPLY-PNH and APPOINT-PNH, Hb levels were measured as part of the primary composite end point (APPLY-PNH primary end points: ≥2 g/dL Hb increase from baseline and Hb ≥12 g/dL, both between days 126 and 168 in the absence of red blood cell transfusion between days 14 and 168; APPOINT-PNH primary end point: ≥2 g/dL Hb increase from baseline in the absence of red blood cell transfusion).¹⁸ Secondary end points examined included change from baseline in Hb, FACIT-Fatigue, and LDH, each assessed using the model-adjusted mean of visits on days 126, 140, 154, and 168.¹⁸ 

The PROs assessed in APPLY-PNH and APPOINT-PNH included the FACIT-Fatigue, EORTC QLQ-C30, and PGIS. These assessments were completed on an electronic device at screening and on days 1, 7, 14, 42, 84, 126, 140, 154, and 168, except for the day 7 visit when only the FACIT-Fatigue and PGIS were collected. The change from baseline in these PRO measures was analyzed using longitudinal models, including all visits, and summarized using the model-adjusted mean of visits on days 126, 140, 154, and 168.

The analyses for the PRO studies were performed based on the PRO-evaluable set, which consisted of patients in the APPLY-PNH or APPONT-PNH full-analysis set with baseline FACIT-Fatigue scores.

The anchor-based determination of the MWPC threshold for FACIT-Fatigue has been previously described in detail.²⁰ Briefly, a 5-point improvement from baseline in FACIT-Fatigue scores was prespecified in APPLY-PNH and APPOINT-PNH²³; however, anchor-based MWPC using the PGIS determined the threshold range to be 7.5 to 9.5 points.²⁰ A 9-point change from baseline, which represented the upper bound of this range, was chosen as a conservative MWPC threshold value for this analysis. A patient was considered to meet the criteria for MWPC if the change from baseline in FACIT-Fatigue score at a set time point was greater than or equal to this set threshold. The standardized marginal proportion of patients who achieved MWPC was averaged over the last 4 visits (days 126, 140, 156, and 168), and comparisons between the treatment groups presented as differences and ratios.

Change from baseline on the EORTC QLQ-C30 mirrored the analysis used for FACIT-Fatigue but focused on subscales that were most relevant for PNH. In brief, the assessment of changes in these subscales was derived using a longitudinal model and the use of anchors to determine the magnitude of MWPC. Change from baseline was analyzed for the physical (items 1-5) and role functioning (items 6-7) subscales; the fatigue (items 10, 12, and 18), dyspnea (item 8), pain (items 9 and 19), and insomnia (item 11) symptom subscales; and the global health status (items 29 and 30) subscale. MWPC thresholds were derived for the physical and role functioning subscales and symptom subscales of fatigue and dyspnea. The reported proportions, differences, and ratios for each subscale were determined by the means of visits on days 126, 140, 154, and 168.

Changes in hematologic markers (Hb and LDH) and patient-reported fatigue (FACIT-Fatigue) were measured using the mean change from baseline as secondary end points in APPLY-PNH and APPOINT-PNH. The correlation between FACIT-Fatigue scores and either Hb levels or LDH was evaluated using the Pearson correlation coefficient. For APPLY-PNH, correlations were calculated using data collected on day 168 only. In APPOINT-PNH, correlations were calculated using data collected at both baseline and day 168.

From APPLY-PNH (N = 97), 95 of 97 C5i-experienced patients with persistent anemia were included in the PRO-evaluable set and randomized to receive iptacopan (n = 62) or C5i therapy (n = 33). Two patients from APPLY-PNH did not provide evaluable PRO data and were excluded from the PRO-evaluable set. All 40 C5i-naive patients who started study treatment in APPOINT-PNH were included in the PRO-evaluable set.

The study population and patient characteristics have been previously described in detail.¹⁸ In APPLY-PNH, the mean baseline Hb was 8.9 g/dL, and 58% received transfusions 6 months before randomization. Mean baseline FACIT-Fatigue scores were 34.7 in the iptacopan group and 30.8 in the C5i group. In APPOINT-PNH, mean baseline Hb, LDH, and FACIT-Fatigue scores were 8.2 g/dL, 1698.8 U/L, and 32.8, respectively, and 70% of patients received transfusions in the 6 months before study treatment (Table 1).

In APPLY-PNH, a higher mean proportion of C5i-experienced patients in the iptacopan group (62%; 95% confidence interval [CI], 55-70]) than the C5i group (29%; 95% CI, 22-35) achieved a 5-point improvement in FACIT-Fatigue scores from baseline (difference: 34%; 95% CI, 29-39). In addition, the MWPC threshold of 9 points was reached by a greater mean proportion of patients treated with iptacopan (51%; 95% CI, 43-59) than with C5i (11%; 95% CI, 5-19). The difference between the marginal proportions for iptacopan vs C5i was 39% (95% CI, 31-49), indicating strong evidence of an improvement in iptacopan compared with C5i-treated patients, and C5i-experienced patients treated with iptacopan were ∼4 times more likely to experience MWPC in fatigue than patients on C5i (Figure 1A). The strength of this result was confirmed with empirical cumulative distribution function curves that depict a clear separation between treatment arms and significance (P ≤ .0039) according to the Mann-Whitney U test at the 9-point threshold favoring iptacopan (Figure 1B).

In APPOINT-PNH, a mean proportion of 68% (95% CI, 56-79) C5i-naive patients treated with iptacopan achieved a 5-point improvement in FACIT-Fatigue scores, and 56% (95% CI, 43-68) achieved the MWPC threshold of 9 points (Figure 1C).

The EORTC QLQ-C30 function subscales indicated that C5i-experienced patients treated with iptacopan (APPLY-PNH) experienced an estimated average improvement from baseline of 14.7 and 17.4 points, respectively, in physical functioning and role functioning compared with patients treated with C5i. A similar result was derived from the symptom subscales, which showed that C5i-experienced patients treated with iptacopan experienced larger improvements in fatigue (−16.8 points), dyspnea (−22.4 points), and pain (−7.4 points) than patients treated with C5i. The global health index showed an increase of 13.8 points in the iptacopan group compared with the C5i group. Improvement from baseline in the insomnia symptom subscale was minimal (−0.9 points), with the CI on both sides of 0 indicating no treatment effect (Table 2). Nonparametric Mann-Whitney U effect measures and P values confirm the strength of improvements for patients treated with iptacopan compared with C5i (supplemental Figure 1).

C5i-naive patients treated with iptacopan (APPOINT-PNH) improved from baseline by an estimated average of 15.2 points (physical subscale), 18.9 points (role subscale), −21.9 points (fatigue subscale), −11.4 points (pain subscale), −15.8 points (dyspnea subscale), and 24.4 points (global health subscale). The improvement from baseline in insomnia was minimal (−5.3 points), with the upper bound of the CI close to 0 (Table 2).

The EORTC QLQ C-30 analysis of MWPC found that C5i-experienced patients (APPLY-PNH) on iptacopan had stronger improvement in the impact of PNH than patients on C5i. The average proportion of patients who achieved the MWPC threshold of 18 points for the physical functioning subscale was 40% (95% CI, 32-48) for iptacopan vs 9% (95% CI, 5-13) for C5i (difference: 31%; 95% CI, 25-38). The role functioning subscale had similar results, with 39% (95% CI, 31-47) of iptacopan-treated patients reaching the 18-point MWPC threshold vs 15% (95% CI, 10-20) of C5i-treated patients (difference: 24%; 95% CI, 19-29). The symptom subscales of fatigue and dyspnea were evaluated using thresholds of −20 and −21, respectively. A higher proportion of patients treated with iptacopan than C5i achieved thresholds for fatigue (49%; 95% CI, 41-56 vs 14%; 95% CI, 9-19) and dyspnea (46%; 95% CI, 38-54 vs 20%; 95% CI, 14-27) (difference, 35%; 95% CI, 30-40 and 26%; 95% CI, 21-31), respectively. C5i-experienced patients treated with iptacopan were 2 to 4 times more likely to experience a MWPC in these EORTC QLQ-C30 subscales than patients on C5i (Figure 2A).

In APPOINT-PNH, anchor-based MWPC thresholds were established for the EORTC QLQ-C30 subscales of physical functioning (18 points), role functioning (22 points), fatigue (−25 points), and dyspnea (−16 points). The average proportion of C5i-naive patients treated with iptacopan who achieved MWPC was 41% (95% CI, 29-53) for physical functioning, 46% (95% CI, 34-58) for role functioning, 46% (95% CI, 34-59) for fatigue, and 55% (95% CI, 42-68) for dyspnea (Figure 2B).

In C5i-experienced patients treated with iptacopan (APPLY-PNH), Hb levels and FACIT-Fatigue scores improved over time.¹⁸ The adjusted least squares mean Hb change from baseline was 3.6 g/dL (95% CI, 3.3-3.9) for patients treated with iptacopan and −0.06 g/dL (95% CI, −0.5 to 0.3) for patients treated with C5i (difference: 3.7 g/dL; 95% CI, 3.2-4.1; P < .001). The adjusted least squares mean change from baseline FACIT-Fatigue score was 8.6 points (95% CI, 6.7-10.5) for patients treated with iptacopan and 0.3 (95% CI, −2.2 to 2.8) for C5i-treated patients (difference: 8.3 points; 95% CI, 5.3-11.3; P < .001). There was no evidence of a treatment difference in the adjusted mean percentage change from baseline in LDH: −3.5% (95% CI, −10.0 to 3.4) for iptacopan vs −2.4% (95% CI, −10.8 to 6.7) for C5i (Figure 3A).¹⁸ In APPLY-PNH, Hb levels correlated with FACIT-Fatigue scores in all patients at day 168. The Pearson correlation value between FACIT-Fatigue score and Hb was 0.48 (P < .001). In contrast, there was no correlation between FACIT-Fatigue score and log-transformed LDH with a Pearson correlation value of −0.15 (P = .16; Figure 3B).

In C5i-naive patients treated with iptacopan (APPOINT-PNH), Hb levels, LDH values, and FACIT-Fatigue scores improved over time.¹⁸ In APPOINT-PNH, the adjusted least square mean Hb change from baseline was 4.3 g/dL (95% CI, 3.9-4.7), the adjusted least square mean change from baseline in FACIT-Fatigue score was 10.8 points (95% CI, 8.7-12.8), and the adjusted mean percentage change from baseline in LDH was −83.6% (95% CI, −84.9 to −82.1), with 95% of patients achieving LDH of ≤1.5 × the upper limit of normal (Figure 3C).¹⁸ In APPOINT-PNH, FACIT-Fatigue scores correlated with both Hb levels and log-transformed LDH at baseline and day 168. The Pearson correlation value was 0.42 (P < .001) for FACIT-Fatigue and Hb and −0.53 (P < .001) for FACIT-Fatigue log-transformed LDH (Figure 3D).

Untreated patients with PNH experience a high symptom burden that has been shown to significantly impact HRQOL and the ability to perform daily activities.^3,25,26 Fatigue is the most commonly reported symptom of PNH and is described by patients with PNH as severe and distressing.^1,3,25,26 Hematologic end points, such as Hb and LDH levels, are excellent for evaluating the clinical efficacy of anticomplement treatment but cannot measure the patient perception of fatigue and other disease-related symptoms that are known to influence HRQOL. Instead, PROs can be used to measure these symptoms and provide important information on the patient’s perspective of their disease and response to treatment.^19,27 In this analysis of APPLY-PNH and APPOINT-PNH, PROs, including the FACIT-Fatigue and EORTC QLQ-C30, were used to assess fatigue and HRQOL in both C5i-experienced and -naive patients treated with iptacopan. The FACIT-Fatigue and EORTC QLQ-C30 are considered relevant for assessing PNH-associated fatigue and HRQOL, respectively, and both have been used extensively in clinical trials and by the International PNH Registry to evaluate quality of life in patients with PNH.^3,19 Baseline FACIT-Fatigue scores were comparable in APPLY-PNH and APPOINT-PNH,¹⁸ indicating that C5i-experienced and -naive patients have similar levels of fatigue and that both intravascular and extravascular hemolysis can contribute to fatigue. The response after receiving treatment with iptacopan, with patients from both studies exhibiting MWPC for the FACIT-Fatigue and EORTC QLQ-C30 questionnaires, also indicates similar unequivocal improvement in overall quality of life after treatment.

There are several ways to interpret PRO data. These outcomes may be measured in absolute terms or with change from a previous measure (ie, change from baseline). To evaluate the clinical meaning and treatment effect of changes in these outcomes, specific thresholds of response have been established, including the clinically important difference (difference between groups of people), clinically important change (change within an individual that is considered clinically meaningful), and MWPC (the amount of change it takes for a patient to feel different).^28,29 The US Food and Drug Administration prefers MWPC and has included it in recent guidance as this measure considers the patient perspective of treatment.³⁰ 

Here, MWPC was assessed using predetermined response thresholds for the FACIT-Fatigue and 4 subscales of the EORTC QLQ-C30 (physical functioning, role functioning, fatigue, and dyspnea). The analysis of the FACIT-Fatigue used 2 separate thresholds: a ≥5-point change from baseline indicating clinical improvement²⁸ and a ≥9-point change from baseline indicating MWPC.²⁰ A higher proportion of C5i-experienced patients treated with iptacopan achieved MWPC compared with patients treated with C5i for both thresholds. However, the treatment effect of iptacopan was more pronounced, with a greater difference between treatments, at the MWPC threshold of 9 points. For the EORTC QLQ-C30, a higher proportion of patients treated with iptacopan showed improvement and reached the MWPC threshold for all 4 subscales compared with patients treated with C5i. These results provide evidence of an additional treatment benefit with iptacopan compared with C5i, with respect to increasing the proportion of patients who experience meaningful improvement in HRQOL. These results are corroborated by the analysis of a semistructured patient interview that focused on the qualitative features of study treatment and disease experience, further confirming the positive experience of treatment and improvement of disease in patients treated with iptacopan vs C5i.³¹ 

Additionally, the relationship between FACIT-Fatigue scores and hematologic parameters of disease (Hb and LDH levels specifically) was investigated in C5i-experienced patients and in C5i-naive patients with Hb of <10 g/dL. In C5i-experienced patients with persistent anemia treated with iptacopan (APPLY-PNH), Hb levels and FACIT-Fatigue scores improved over time,¹⁸ and increased Hb levels were correlated with improved FACIT-Fatigue scores. The temporal association of these changes was also similar, with increased FACIT-Fatigue scores after increased Hb levels from baseline to day 168 (Figure 3A).¹⁸ In C5i-naive patients treated with iptacopan (APPOINT-PNH), Hb, LDH, and FACIT-Fatigue scores improved over time (Figure 3C).¹⁸ In these patients, both increased Hb levels and decreased LDH levels were correlated with improved FACIT-Fatigue scores.

Studies have shown that despite treatment with C5is, many patients experience persistent anemia and fatigue, and have remaining transfusion needs.^8,12-15 It is thought that this is mainly because of emerging extravascular hemolysis,⁸ but the association between extravascular hemolysis and fatigue is unclear. Among patients with cancer and anemia, increased Hb levels are associated with clinically significant improvement in fatigue, particularly when Hb increases from 11 to 12 g/dL.³² However, the validity and dynamics of this relationship in patients with PNH is unknown. A study of complement inhibitor-naive patients with PNH treated with ravulizumab or eculizumab found that improved LDH (an indicator of reduced intravascular hemolysis) was significantly correlated with improved FACIT-Fatigue and EORTC QLQ-C30 global health scores despite nonsignificant changes in Hb levels.³³ Similar to what was observed in APPLY-PNH, an analysis of C5i-experienced patients with persistent anemia treated with pegcetacoplan reported that improved FACIT-Fatigue scores were correlated with increased Hb levels (an indicator of reduced all-cause hemolysis).³⁴ This study also found that markers of extravascular hemolysis such as reticulocyte count and indirect bilirubin improved along with FACIT-Fatigue scores but failed to mention LDH.³⁴ Here, we show that both relationships are true in specific patient populations. In C5i-experienced patients treated with iptacopan, improved FACIT-Fatigue scores were correlated with improved Hb levels but not LDH. In C5i-naive patients treated with iptacopan, improved FACIT-Fatigue scores were correlated with both increased Hb levels and decreased LDH levels. Therefore, fatigue in patients with PNH may have 2 sources, intravascular hemolysis and lower-than-normal Hb levels. Interventions that increase Hb levels by addressing both intravascular and extravascular hemolysis may be important for decreasing the symptoms and experience of fatigue in all patients with PNH.

Some limitations of this study should be considered. Both APPLY-PNH and APPOINT-PNH were open-label study designs, which may influence their perspectives on treatment and treatment success. Trials investigating rare diseases like PNH are also subject to a small sample size.^18,35,36 However, despite the small sample size in APPLY-PNH, a clear difference between the treatment groups was estimated with good precision for both the FACIT-Fatigue and EORTC QLQ-C30 subscales, and a nonparametric Mann-Whitney U test supported these results. PROs are subject to the potential for patient anticipation of improvements because of the initiation of new therapy, especially with prespecified study visits for assessment and an open-label study design, which may introduce bias and influence results.³⁷ Patients on iptacopan reported consistent improvements in PRO scores, and the changes experienced were both large and durable, dispelling the possibility that the results were based on bias alone. Additionally, the time course of improvement in the PRO measures with little change recorded 7 days after randomization, reinforces the conclusion that patients were recording perceived improvement rather than bias because of the open-label study design. Both the FACIT-Fatigue and EORTC QLQ-C30 were designed to evaluate HRQOL in patients with cancer.³⁸ Although these PROs have strong support for their validity for use in PNH and are frequently used in the field,¹⁹ other PROs have been specifically designed to assess the symptoms and experience of PNH and could be used in the future to increase the specificity of these results.^38-40 

Analysis of MWPC for the FACIT-Fatigue and 4 EORTC QLQ-C30 subscales using patient-centered anchors found that both C5i-experienced patients with persistent anemia and C5i-naive patients experienced meaningful improvements in the symptoms and impacts of PNH during treatment with iptacopan. Improvements in FACIT-Fatigue scores were associated with increased Hb levels in C5i-experienced patients and with both increased Hb levels and decreased LDH levels in C5i-naive patients treated with iptacopan. These results, combined with efficacy data from APPLY-PNH and APPOINT-PNH, indicate that iptacopan results in clinically meaningful disease control and meaningful improvements in the fatigue and HRQOL impacts of PNH after 168 days of treatment in C5i-experienced and -naive patients with PNH.

The authors thank the patients and the investigators who were a part of these trials. Outputs on the correlation among lactate dehydrogenase, hemoglobin, and fatigue were produced by Ranjan Tiwari of Novartis Pharmaceuticals Corporation. Editorial and medical writing support was provided by Erika Tomei of Nucleus Global and was funded by Novartis Pharmaceuticals Corporation.

This study was funded by Novartis Pharmaceuticals Corporation.

This manuscript was developed in accordance with Good Publication Practice (GPP 2022) guidelines.

Contribution: A.M.R., C.d.C., B.H., A.K., J.P.M., P.S., Y.U., and R.P.d.L. participated in conceptualization, data collection, data analysis and interpretation, drafting of the article, clinical revision of the article, and final approval of the published version; S.V., G.B., M.D., and R.K. participated in conceptualization, data analysis and interpretation, drafting of the article, clinical revision of the article, and final approval of the published version; all authors had full control of the content and made the final decision on all aspects of this publication.

Conflict-of-interest disclosure: A.M.R. reports research support from Alexion, Novartis, Alnylam, and Ra Pharma; consultancy from Amyndas; speakers fees from Alexion, Novartis, Pfizer, and Sobi; and membership on an entity’s board of directors or advisory committees of Alexion, Apellis, Novartis, F. Hoffmann-La Roche, Achillion, and Samsung. C.d.C. reports consulting fees from Alexion and Apellis; honoraria from BioCryst, Omeros, and Regeneron; speakers fees from Alexion, Apellis, and Novartis; and membership on an entity’s board of directors or advisory committees of Novartis. A.K. reports research support from Novartis and Bristol Myers Squibb; consulting fees from Alexion, Novartis, Amgen, Agios, Pfizer, Samsung, Celgene, F. Hoffmann-La Roche, and Sobi; honoraria from Alexion, Novartis, Pfizer, Amgen, Samsung, Celgene, F. Hoffmann-La Roche, Bristol Myers Squibb, Sobi, and Silence Therapeutics; and speakers fees from Alexion, Novartis, Amgen, Pfizer, Celgene, F. Hoffmann-La Roche, Sobi, and Ra Pharma. J.P.M. reports consulting fees from Regeneron and Omeros; honoraria from Novartis and Regeneron; speakers fees from Novartis; and membership on an entity’s board of directors or advisory committees of Alexion. P.S. reports research support from Alnylam, BioCryst, Novartis, Pfizer, and F. Hoffmann-La Roche; consulting fees from Alexion, AstraZeneca, BioCryst, Janssen, Novartis, Pfizer, and F. Hoffmann-La Roche; and speakers fees from Alexion, AstraZeneca, Bristol Myers Squibb, Janssen, Novartis, Pfizer, F. Hoffmann-La Roche, and Novartis. Y.U. reports research support from Chugai; consulting fees from Alexion, Asahi Kase, Chugai, Janssen, Novartis, Ono, Sanofi, and Sobi; honoraria from Alexion, Chugai, Kaken, Nippon Shinyaku, Sobi, Incyte, and Sanofi; speakers fees from Alexion, Novartis, and Sanofi; and membership on an entity’s board of directors or advisory committees of Alexion, Janssen, Novartis, and Sanofi. S.V. is an employee of Novartis Services Incorporated and holds shares and/or stock options in the company. G.B. was an employee of Novartis Pharma AG. M.D. is an employee of Novartis Pharma AG and holds shares and/or stock options in the company. R.K. is an employee of Novartis Healthcare Private Limited and holds shares and/or stock options in the company. R.P.d.L. reports research support from Alexion, Pfizer, Novartis, Jazz, and Amgen; consulting fees from Alexion, Pfizer, Novartis, Sobi, Roche, Samsung, Keocyt, Merck Sharp & Dohme, Gilead, Jazz, and Amgen; and honoraria from Alexion, Pfizer, Novartis, Sobi, Roche, Samsung, Keocyt, Merck Sharp & Dohme, Gilead, Jazz, and Amgen. B.H. declares no competing financial interests.

Correspondence: Susan Vallow, Novartis Services Inc, 1 Health Plz, East Hanover, NJ 07936; email: susan.vallow@novartis.com.