Abstract
BACKGROUND: Paroxysmal nocturnal hemoglobinuria (PNH) is a rare, genetic, chronic, and life-threatening blood disease with an estimated prevalence of 13 per 1,000,000 persons reported in the United States. Available at analysis, PNH treatment included the use of C5 inhibitors (C5is), which prevent formation of membrane attack complex and consequently intravascular hemolysis. Limited real-world evidence suggests some individuals with PNH continue to experience anemia and breakthrough hemolysis (BTH) after C5i treatment, indicating unmet needs.
OBJECTIVE: To describe real-world treatment patterns and outcomes among individuals treated with C5is, eculizumab (ECU), and ravulizumab (RAV), focusing on affordability challenges and therapy unmet needs from a US payer perspective.
METHODS: This retrospective cohort study was conducted using deidentified data from Prime Therapeutics’ approximately 15 million commercially insured US members with integrated medical and pharmacy claims data. Members were identified between January 1, 2018, and December 31, 2020. Inclusion criteria for cohort identification were adults aged 18 years or older at ECU or RAV index date requiring 2 or more claims for ECU or 1 or more claims for RAV. ECU and RAV users were excluded if they had a claim indicating treatment for a US Food and Drug Administration (FDA)–approved non-PNH indication. Members were required to be continuously enrolled 6 months before and 12 months after their index ECU or RAV claim. Real-world C5i claims-based treatment dosage and frequency patterns were compared with FDA-labeled dosing. Clinical outcomes, including transfusions and BTH events, were identified in the pre-index and post-index periods. Health care resource use and costs were calculated after network discounts, including member share.
RESULTS: A total of 86 commercial members met analysis criteria: 34 in the ECU cohort and 52 in the RAV cohort. The mean age was 42.6 years, and 54.6% were female. Estimated higher-than-label PNH-recommended dosage occurred in 38.2% of ECU and 9.6% of RAV members. In total, 29.4% of ECU and 17.3% of RAV members had 4 or more transfusions in the post-index period. Additionally, 29.4% of ECU and 13.5% of RAV members had 1 or more BTH episodes. Post-index period mean per member total health care costs were $711,785 among ECU members and $624,911 among RAV members, and C5i costs accounted for 79.7% and 85.6% of total health care costs, respectively.
CONCLUSIONS: Although all members received at minimum FDA-approved dosages, transfusions and BTH events continue to occur for some members. These findings indicate potentially inadequate therapy responses in a substantial subset of C5i users, adding additional therapy costs to an already extremely expensive therapy.
DISCLOSURES: This study was funded by Apellis Pharmaceuticals. Drs Broderick and Fishman report employment by Apellis Pharmaceuticals and own stock options. Dr Burke reports employment by Prime Therapeutics, LLC, which has received research funding from Apellis Pharmaceuticals. Dr Gleason reports employment by Prime Therapeutics, LLC, which has received research funding from Apellis Pharmaceuticals; serves on the advisory committee at the Institute for Clinical and Economic Review; and has served on the Board of Directors at the Academy of Managed Care Pharmacy.
Plain language summary
In this real-world study of members treated with eculizumab (ECU) and ravulizumab (RAV) (C5-inhibitors [C5is]) for paroxysmal nocturnal hemoglobinuria (PNH), discontinuation was observed in one-third of members, and a subset of members did not appear to have their condition adequately controlled as they experienced dose escalation above the US Food and Drug Administration–approved dose, breakthrough hemolysis, and/or remained transfusion-dependent. C5i drug costs were more than 75% of the PNH total cost of care, exceeding an average of $500,000 in the first year of therapy.
Implications for managed care pharmacy
Given the high total cost of care associated with the C5i medications ECU and RAV for PNH, an orphan/rare disease, this analysis may have implications for the development of ongoing PNH drug therapy management surveillance to confirm need for dose escalation and ensure outcomes are being achieved. A transfusion-free outcome may confirm clinical benefits are received from ECU and RAV, while maintaining affordability and ensuring PNH is well controlled.
Paroxysmal nocturnal hemoglobinuria (PNH) is a rare (US prevalence = 13 per 1,000,0001,2), genetic, chronic, and life-threatening blood disease,2 which results in the production of red blood cells (RBCs) that lack complement regulatory proteins required for proper immune function.2 Because of their abnormal complement system, the RBCs in patients with PNH are susceptible to 2 different types of hemolysis via the complement system: intravascular and extravascular.2,3 Intravascular hemolysis occurs in the circulatory system because of the unregulated formation of membrane attack complexes (MACs).2-4 MAC forms holes in the cell membrane and causes them to lyse.4,5 Conversely, extravascular hemolysis occurs outside the circulatory system and is caused by unregulated complement-related opsonization of RBCs, which targets them for destruction in the spleen or liver.2-5 Treating both types of hemolysis is important for PNH disease control in patients who are susceptible to life-threatening thrombotic events and severe, debilitating anemia.3,6
Orphan drugs are developed to treat rare diseases, affecting less than 200,000 individuals in the United States.7 The number of medications to treat rare diseases has grown substantially from 37% of total US Food and Drug Administration (FDA) approvals in 2011 to 52% in 2021.8,9 Although these drugs treat diseases with limited or no available treatment, they are associated with 27-fold higher annual costs, compared with nonorphan drugs.10 Given the increase in approved orphan drugs, this analysis sought to evaluate 1 indication within 1 category of orphan drugs, complement inhibitors, given their higher costs and the paucity of evidence that exists on the real-world use among 2 of the 3 approved treatments for PNH, eculizumab (ECU), and ravulizumab (RAV).
Historically, transfusions, which can temporarily improve anemia-related symptoms, were used for treating hemolysis or anemia secondary to PNH; however, refractoriness to transfusions can develop.11,12 Recently, C5 inhibitor (C5i) treatments were considered the standard treatment following the approval of the monoclonal antibody ECU by the FDA in 2007.11 Ravulizumab (RAV) is similar to ECU and was FDA approved in 2018 based on 2 phase 3 noninferiority trials comparing RAV to ECU (Study 301, n = 246; Study 302, n = 195).11,13,14 Both C5i treatments prevent formation of MAC and intravascular hemolysis; however, they may not be effective in some patients with PNH.11 Specifically, they do not address anemia associated with extravascular hemolysis occurring in most treated patients with PNH.15 Additionally, some patients with PNH experience a return of intravascular hemolysis and PNH symptoms, which C5is should mechanistically control.11,16,17 The explanation may be related to disease complexity, as breakthrough hemolysis (BTH) or anemia following C5i treatment may arise from incomplete complement inhibition or from complement-amplifying conditions, such as infection, surgery, or pregnancy overriding the C5i treatment.17 The subsequent poor disease control in PNH may require using higher-than-label C5i doses and blood transfusions to counteract insufficient complement inhibition.11
With the expanded treatments available to treat PNH,18 it is of increasing importance to understand the real-world use and outcomes of these medications, which may assist health plans with their clinical and economic evaluation of the PNH treatment category. Outcome metrics that may be considered as criteria to confirm clinical benefit or to confirm the treatment is achieving satisfactory control of the disease can be identified through real-world treatment pattern studies.
The analytic goal of this work was to describe real-world treatment patterns and outcomes, including transfusions, BTH events, health care resource use, and costs of individuals treated with C5i treatments, ECU and RAV, focusing on understanding affordability and unmet needs from a US payer perspective. Because of the rare nature of PNH and the paucity of real-world data available on the C5i treatment patterns, this analysis was primarily descriptive and did not hypothesize the presence of a specific treatment effect or magnitude of effect.
Methods
STUDY DESIGN AND DATA SOURCE
This retrospective cohort analysis was conducted using deidentified data from Prime Therapeutics’ commercial book of business. Members were identified between January 1, 2018, and December 31, 2020, which is prior to the FDA approval of the C3/C3b inhibitor, pegcetacoplan. Prime’s database consists of eligibility records, medical claims, and pharmacy claims data for approximately 15 million commercially insured members across the United States. Because this analysis was done with the goal of improving managed care pharmacy business services, institutional review board approval was not required.
MEMBER SELECTION
Medical codes from the Healthcare Common Procedure Coding System (HCPCS) were used to identify members treated with ECU or RAV. Inclusion criteria included members who were aged 18 years or older at the index date and had either 2 or more medical or pharmacy claims for ECU (HCPCS = J1300; Medi-Span generic product identifier = 85200050002020) or 1 or more medical or pharmacy claims for RAV (HCPCS = J1303 or C9052; Medispan generic product identifier: 85800080202020, 85800080202045, 85800080202060). Of note, only 1 claim was required to maximize the sample size for RAV. Overall, 2 (3.8%) of 52 members in the final analyzable RAV cohort had 1 RAV claim, and all others had at least 2 claims. The index date was defined as the date of first ECU or RAV infusion. To ensure that an assessment of relevant cost outcomes was evaluated in the cohort, members were required to have 182 days of continuous enrollment prior to the index date, used to define the pre-index period, as well as 365 days after the index date, which was used to define the post-index period. The overall observation period for each member spanned from the beginning of the pre-index period to the end of the post-index period. Consistent with other retrospective analyses in PNH, we sought to define the population by an exclusion principle rather than rely on International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10)-based diagnoses for PNH, which can be imprecise.19,20 Using this principle, members were excluded if they had 1 or more diagnoses of another indication of ECU or RAV, rather than a conventional approach of including members with 1 or more diagnoses of PNH, helping to improve specificity of the treated population and ensure that both medications were given for the indication of PNH. Additionally, sensitivity analyses were performed among a subset of members in the sample who had 2 or more medical claims with an ICD-10-CM code for PNH in the pre-index or post-index periods. Members were excluded if they had 1 or more diagnoses of another indication for ECU during the pre-index period or on the index date as identified based on diagnosis codes from the ICD-10-CM that included atypical hemolytic uremic syndrome (ICD-10-CM = D59.3), generalized myasthenia gravis (ICD-10-CM = G70.0x), and/or neuromyelitis optica spectrum disorder (ICD-10-CM = G36.0). This constituted the overall population cohort (for all HCPCS and ICD-10-CM codes, see Supplementary Table 1, available in online article). To define each of the 2 treatment cohorts within the overall population cohort, those members with claims for both ECU and RAV in the identification period were included in the RAV cohort, and their first claim for RAV was considered their index claim. Members who only received ECU during the entire observation period were included in the ECU cohort.
MEMBER CHARACTERISTICS
Age, sex, and Quan-Charlson comorbidity index score were determined on the index date (Table 1). Clinical characteristics, including aplastic anemia, anemia, bleeding, myelodysplastic syndrome, infection, thrombosis, and dyspnea, were identified in the pre-index and post-index periods. For each characteristic, members were required to have 2 or more medical claims with an ICD-10-CM code in any position 30 days or more apart to be defined as having the condition in the pre-index or post-index period. BTH was defined with coded symptoms or complications commonly used to define BTH and intravascular hemolysis in PNH clinical trials and clinician guidance (see list of codes in Supplemental Material).13,14,17,21-23 All codes for BTH on a single day were considered a single episode. However, if a BTH was identified via BTH symptoms and a hospitalization, the BTH episode was extended through the entire length of the hospitalization.
TABLE 1.
PNH Cohort, Patient Characteristics, by Index C5i Therapy
| Variable | Eculizumab(n = 34) | Ravulizumab(n = 52) |
|---|---|---|
| Baseline characteristics | ||
| Age, mean (SD) | 40.9 (12) | 43.7 (13.8) |
| Age group, n (%) | ||
| 18-49 years | 25 (73.5) | 34 (65.4) |
| 50-64 years | 8 (23.5) | 15 (28.9) |
| ≥ 65 years | 1 (2.9) | 3 (5.8) |
| Sex, n (%) | ||
| Female | 18 (52.9) | 29 (55.8) |
| Male | 16 (47.1) | 23 (44.2) |
| Quan-Charlson comorbidity index, mean (SD) | 2.4 (2.7) | 1.5 (2.7) |
| HRU and costs in pre-index period | ||
| Total HRU cost of care, mean (SD) | $263,665 ($208,729) | $240,993 ($153,667) |
| Total HRU cost of care, median | $251,034 | $242,809 |
| Number of inpatient visits, mean (SD) | 2.1 (1.6) | 1.6 (0.8) |
| Number of ED visits, mean (SD) | 1.3 (0.5) | 1.6 (1.2) |
| Number of transfusions, mean (SD) | 7.4 (8.1) | 7.4 (7.0) |
| Pre-index and post-index comorbidities and clinical complications, n (%) | ||
| Aplastic anemia | ||
| Pre-index period | 12 (35.3) | 23 (44.2) |
| Post-index period | 17 (50.0) | 23 (44.2) |
| Anemia | ||
| Pre-index period | 21 (61.8) | 34 (65.4) |
| Post-index period | 23 (67.7) | 34 (65.4) |
| Bleeding | ||
| Pre-index period | 0 (0.0) | 3 (5.8) |
| Post-index period | 3 (8.8) | 3 (5.8) |
| BTH | ||
| Pre-index period | 8 (23.5) | 7 (13.5) |
| Post-index period | 10 (29.4) | 7 (13.5) |
| MDS | ||
| Pre-index period | 4 (11.8) | 4 (7.7) |
| Post-index period | 4 (11.8) | 5 (9.6) |
| Infection | ||
| Pre-index period | 9 (26.5) | 6 (11.5) |
| Post-index period | 12 (35.3) | 14 (26.9) |
| Thrombosis | ||
| Pre-index period | 1 (2.9) | 4 (7.7) |
| Post-index period | 5 (14.7) | 7 (13.5) |
| Dyspnea | ||
| Pre-index period | 12 (35.3) | 9 (17.3) |
| Post-index period | 9 (26.5) | 10 (19.2) |
All costs are shown in US dollars.
BTH = breakthrough hemolysis; C5i = C5 inhibitor; ED = emergency department; HRU = health care resource utilization; MDS = myelodysplastic syndrome; PNH = paroxysmal nocturnal hemoglobinuria.
Comorbidities, including aplastic anemia and myelodysplastic syndrome, were identified in the pre-index and post-index periods among members with 2 or more medical claims with an ICD-10-CM code in any position at least 30 days apart (see list of codes in Supplemental Material). The Quan-Charlson comorbidity index was also used to categorize comorbidities.24
TREATMENT PATTERN OUTCOMES
Real-world treatment patterns were captured, including treatment duration, discontinuation, switching (post-index period), previous ECU treatment (pre-index period), dosage (units of drug administered), and frequency (time between claims). Treatment duration was characterized by the number of days a member was on ECU or RAV to discontinuation, switching, or end of study period. For members who switched treatment, duration of therapy for the second therapy was also determined. Discontinuation was characterized as a gap in index therapy in the post-index period. Lacking a validated algorithm to define discontinuations, a gap of at least 2 times the approved dosing interval for each product was selected. As such, a member was considered to have discontinued therapy if there were no claims in the last 28 or 112 days of the post-index period for ECU and RAV, respectively.
When analyzing treatment patterns based on dosage, members were classified as receiving higher, normal, or lower than standard of care treatment, according to the FDA-labeled dosing. An assessment of the timing of dose escalation in the post-index period was performed by analyzing which members increased their dose relative to the first quarter in each of the three quarters after they initiated their index treatment (Table 2). Dose escalation was defined after members reached maintenance dosing. Normalizing the dose over time accounted for batch billing. Among members new to therapy, the maintenance dose was established after accounting for initial loading doses (ECU = 5; RAV = 1). The Number of Services field in the medical claims data provided the number of medication units. Daily maintenance dose was determined by dividing the total number of medication units during maintenance therapy by the number of days on maintenance therapy. This categorization was limited for RAV, which is weight-based. Because the database does not include member weight, the threshold for RAV was set using the highest labeled dose and interval (3,600 mg every 56 days). Estimating or using an approximate dose for weight-based intravenous drugs is a common occurrence in claims analyses when weight-based data are unavailable.25-28 Members with a 25% greater or less than FDA-approved maintenance dose were defined as not having a standard of care treatment.
TABLE 2.
Eculizumab and Ravulizumab Treatment Patterns in the 12-Month Post-Index Period
| Outcome | Eculizumab (n = 34) | Ravulizumab (n = 52) |
|---|---|---|
| New to all C5i therapy on index date, n (%) | 16 (47.2) | 17 (32.7) |
| New to index therapy on index date, n (%) | 16 (47.2) | 52 (100) |
| Discontinuation | ||
| Discontinuation of therapy, n (%)a | 10 (29.4) | 17 (32.7) |
| PNH ICD-10-CM subgroupb: discontinuation of therapy, n (%)a | 4 (17.4) | 13 (27.7) |
| Mean days to discontinuation, mean (SD) | 119.0 (67.9) | 274.9 (100.3) |
| PNH ICD-10-CM subgroupb: mean days to discontinuation, mean (SD) | 154.8 (84.6) | 284.5 (96.6) |
| Dose administered as maintenance therapy (excludes loading dose), n (%) | ||
| Normal (dosing per label recommendation) | 16 (47.0) | 43 (82.7) |
| Higher than standard | 13 (38.2) | 5 (9.6) |
| Lower than standard | 0 (0.0) | 0 (0.0) |
| Unknownc | 5 (14.7) | 4 (7.7) |
| Any dose escalation, 25% increase,d n (%) | ||
| Dose escalation at any point Q2-Q4 after Q1 | 1 (2.9) | 16 (30.8) |
| Dose escalation Q2 | 1 (2.9) | 5 (9.6) |
| Dose escalation Q3 | 0 | 12 (23.1) |
| Dose escalation Q4 | 0 | 13 (25.0) |
| Administration of therapy locations, n (%) | ||
| Hospital outpatient | 5 (14.7) | 21 (40.4) |
| Office | 19 (55.9) | 24 (46.2) |
| ED | 0 (0.0) | 0 (0.0) |
| In-home | 2 (5.9) | 4 (7.7) |
| Other | 8 (23.5) | 3 (5.8) |
a No claims for the index therapy in the 12-month post-index period after the old discontinuation date.
b A sensitivity analysis was performed among the subgroup of members in the sample with confirmed PNH diagnosis based on at least 2 medical claims with an ICD-10-CM code for PNH in the pre-index or post-index period (eculizumab: n = 23; ravulizumab: n = 47).
c Five members in the eculizumab cohort and 4 members in the ravulizumab cohort did not reach maintenance dosing and/or switched therapy.
d Members with a 25% greater- or less-than expected milligrams per day during their maintenance period were defined as not having a standard of care treatment. By normalizing the dose over time, batch billing was accounted for. For members new to therapy, the maintenance dose was established after accounting for the initial loading doses (5 for eculizumab and 1 for ravulizumab). The “number of services” variable in the medical claims data provided the number of units of medication. The daily maintenance dose was determined by dividing the total number of units of medication while on maintenance therapy by the number of days on maintenance therapy.
C5i = C5 inhibitor; ED = emergency department; ICD-10-CM = International Classification of Diseases, Tenth Revision, Clinical Modification; PNH = paroxysmal nocturnal hemoglobinuria; Q = quarter.
Frequency of treatment was characterized by the time between claims; expected intervals for members treated with ECU and RAV were 14 and 56 days, respectively. These intervals were included in the analysis to provide an alternate definition of dose increase that was independent of weight-based dosing.
The number of transfusion episodes was determined in the post-index period, and percentage with 0, 1-3, and 4 or more transfusions in the post-index period were reported. All transfusions on a single day were considered a single episode. Time from the index date to the first transfusion and the time between the first and second transfusions were calculated.
HEALTH CARE UTILIZATION AND COST OUTCOMES
Hospitalizations, emergency department (ED), and office visits were captured; PNH-related health care utilization included medical claims with ICD-10-CM code for PNH in any position. In addition to resource use, all-cause and PNH-related costs were captured. The total cost of care included all pharmacy and medical claims, including out-of-pocket costs and plan-paid allowed amounts after network discounts. Costs were broken out by medical costs and pharmacy costs. Medical costs were further broken out by hospitalizations, ED, office visits, and other costs using place of service and revenue codes. PNH-related health care costs included medical claims with an ICD-10-CM code for PNH in any position, as well as costs related to PNH treatment.
STATISTICAL ANALYSES
Descriptive statistical analyses were used instead of any modeling because of sample size limitations. The mean and SD were calculated for continuous variables. Numbers and proportions were calculated for categorical variables. The expected population for this descriptive analysis was at least 60, given PNH is an ultra-rare disease. Analyses were conducted using SAS software version 9.4 m6 (SAS Institute, Inc.).
Results
MEMBER AND CLINICAL CHARACTERISTICS
Between January 1, 2018, and December 31, 2020, 452 members were identified with 2 or more claims of ECU or 1 or more claims of RAV. Of these members, 86 (19.0%) met the inclusion and exclusion criteria (ECU cohort, n = 34; RAV cohort, n = 52) (Figure 1). The mean [SD] age (40.9 [12] and 43.7 [13.8], ECU and RAV, respectively) and distribution of age groups were similar for both cohorts (Table 1). Member sex was also similar with an approximate 1:1 ratio of male and female members for both cohorts. The mean Quan-Charlson comorbidity index score was 2.4 (2.7) in the ECU cohort and 1.5 (2.7) in the RAV cohort. The mean total cost of care and number of inpatient visits were slightly numerically higher for members treated with ECU vs RAV prior to the index date ($263,665 [$208,729] and $240,993 [$153,667], respectively). Both cohorts, however, had a similar frequency of transfusions prior to the index period. The burden of comorbidities was also similar for both cohorts. Pre-index and post-index clinical characteristics, such as aplastic anemia, and complications, such as infections, thrombosis, and bleeding/BTH, are also shown in Table 1.
FIGURE 1.
Cohort Identification Attrition
TREATMENT PATTERNS OVER 12 MONTHS
Of all members receiving the index treatment, the proportion of members naive to C5i treatments on the index date were 47.2% and 32.7% for the ECU and RAV cohorts, respectively (Table 2). In the RAV cohort, 35 members (67.3%) had a claim for ECU in the pre-index period. The proportion of members that were newly initiating their index therapy was 47.2% and 100% for ECU and RAV, respectively. In total, 29.4% of members receiving ECU and 32.7% of RAV members discontinued treatment. A subgroup analysis was performed among those who discontinued to evaluate any temporal association of transfusions as an explanation for discontinuation. A substantial proportion of members discontinuing ECU and RAV required a transfusion prior their discontinuation date (46% and 57%, respectively).
When evaluating dosing for both treatments, the maintenance dose exceeded the FDA recommended dose in 38.2% of members treated with ECU and 9.6% of members treated with RAV (Table 2). When analyzing time between claims for ECU, almost 15% of members in the ECU cohort received treatment more frequently than expected (time between claims: 5.9% at ≤ 9, 0% at 10-11, 8.8% at 12-13, 70.6% at ≥ 14 days, and 14.7% unknown due to not reaching maintenance dosing, respectively). For the RAV cohort, 21.2% were receiving treatment more frequently than expected (0% at < 42, 0% at 42-48, 21.2% at 49-55, 71.1% at > 56 days, and 7.7% unknown as above, respectively). Dose escalation occurred most often in the third and fourth quarters for RAV, indicating that the higher dose may be due to lack of disease control, as the loading dose would not be required in quarters 3 or 4 (3-6 months) after initiating therapy. Treatment administration occurred most commonly in an office setting for both ECU and RAV. In-home administration was relatively uncommon for both treatments.
Despite receiving a C5i treatment, BTH and regular transfusions were relatively common for both cohorts (Table 3). Of the members receiving ECU, 29.4% experienced at least 1 episode of BTH, whereas 13.5% of the RAV cohort experienced at least 1 episode. Furthermore, transfusions occurred in 52.9% of the ECU cohort and 32.7% of the RAV cohort.
TABLE 3.
Transfusions and BTH Episodes in the 12-Month Post-Index Period
| Outcome | Eculizumab (n = 34) | Ravulizumab (n = 52) |
|---|---|---|
| Members with transfusions, n (%) | ||
| No transfusions | 16 (47.1) | 35 (67.3) |
| 1-3 transfusions | 8 (23.5) | 8 (15.4) |
| ≥ 4 transfusion | 10 (29.4) | 9 (17.3) |
| Members with BTH | ||
| No episodes, n (%) | 24 (70.6) | 45 (86.5) |
| 1 episode, n (%) | 4 (11.8) | 2 (3.8) |
| 2 episodes, n (%) | 2 (5.9) | 2 (3.8) |
| ≥ 3 episodes, n (%) | 4 (11.8) | 3 (5.8) |
| Number of BTH episodes, mean (SD) | 0.9 (2.2) | 0.5 (2.0) |
| Members with ED visit | ||
| ≥ 1 visit, n (%) | 13 (38.2) | 12 (23.1) |
| Number of ED visits, mean (SD) | 1.0 (1.6) | 0.7 (1.8) |
| Members with hospitalization/inpatient stay | ||
| ≥ 1 visit, n (%) | 13 (38.2) | 11 (21.2) |
| Number of hospitalizations, mean (SD) | 1.1 (2.1) | 0.5 (1.1) |
BTH = breakthrough hemolysis; ED = emergency department.
HEALTH CARE COSTS AND UTILIZATION DURING THE 12-MONTH POST-INDEX PERIOD
Members treated with ECU were shown to have a mean total annual cost of $711,785, whereas members treated with RAV had mean annual costs of $624,911 (Table 4). The C5i pharmacy costs accounted for most of the total health care costs through the 12-month period for ECU (79.7% of total cost) and RAV (85.6% of total cost). The mean annual RAV cost was $32,074 (5.7%) lower at $535,439, in contrast to ECU at $567,258. The mean total C5i drug costs were highest in the office setting for the ECU cohort, whereas the RAV cohort costs were highest in the hospital outpatient as 40.4% received RAV therapy in that setting.
TABLE 4.
All-Cause Health Care Costs in the 12-Month Post-Index Period
| All-cause costs | Eculizumab (n = 34),mean (SD) | Ravulizumab (n = 52),mean (SD) |
|---|---|---|
| Medical benefit costs | 695,487 (341,176) | 597,895 (53,460) |
| Hospital | 67,127 (133,713) | 28,124 (81,357) |
| ED | 786 (1,413) | 336 (1,051) |
| Office | 298,977 (306,997) | 209,527 (277,063) |
| Outpatient | 209,281 (384,410) | 254,568 (428,991) |
| Other | 119,315 (236,368) | 105,339 (196,991) |
| PNH C5i drug costs | 567,258 (310,889) | 535,184 (358,439) |
| Pharmacy benefit costs | 16,298 (28,733) | 27,016 (75,189) |
| PNH C5i treatment drug costs, by setting of care | ||
| Office | 294,290 (310,129) | 204,758 (273,814) |
| Hospital outpatient | 66,811 (170,596) | 80,694 (163,931) |
| Home infusion | 49,190 (114,952) | 225,882 (418,478) |
| Other | 156,968 (381,183) | 23,850 (98,389) |
| C5i infusion-related costs | 8,638 (13,869) | 2,560 (2,481) |
| PNH-related treatment costs | 97 (215) | 45 (170) |
| Transfusion costs | 13,441 (29,400) | 7,792 (23,973) |
| Total costs | 711,785 (349,750) | 624,911 (352,284) |
All costs are shown in US dollars.
C5i = C5 inhibitor; ED = emergency department; PNH = paroxysmal nocturnal hemoglobinuria.
The proportions who were hospitalized, visited the ED, or visited a doctor’s office were 38.2%, 38.2%, and 97.1%, respectively, for the ECU cohort and 21.2%, 23.1%, and 98.1%, respectively, in the RAV cohort.
Discussion
The present study examined treatment patterns and associated health care costs in members with PNH receiving the C5i treatments, ECU, and RAV. As PNH is a rare disease, the final analytic sample fell within our expected sample size. Further, this cohort is a similar size to that of populations in phase 3 PNH studies that were the basis for FDA approval of these treatments. Most US payers looking to assess PNH in their membership likely do not have an adequate number of members to perform an analysis, thus this study provides an opportunity to describe the burden of this ultra-rare condition from a payer perspective.
We found that 20% of members receiving C5i treatments appear to be at a higher dose than the standard dose approved by the FDA, and 19% of members were receiving treatments more frequently than recommended by the FDA. Overall, 40.7% of members receiving C5i treatments required at least 1 transfusion within the post-index period (ECU = 52.9%; RAV = 32.7%), and 22.1% of members received 4 or more transfusions throughout the year (ECU = 29.4%; RAV = 17.3%). Furthermore, approximately 20% of members receiving C5i treatments experienced at least 1 episode of BTH. These negative outcomes indicate a potentially inadequate therapeutic response in a substantial subset of members receiving C5i treatments.
A retrospective study of 93 patients with PNH treated with ECU found that only 10% of patients showed a complete response, defined as no transfusion requirement with normal stable hemoglobin level and no evidence of hemolysis, over 12 months of treatment.29 Furthermore, more than one-third of patients in this study required blood transfusions, indicating poor disease control despite treatment with ECU.29 A previous analysis of this same commercially insured population was conducted before RAV was widely available and reported a substantial proportion of patients treated with C5i treatments (ECU, n = 57; RAV, n = 6) were still burdened by BTH and probable anemia, with 39.6% of patients receiving blood transfusions.30 This analysis also demonstrated that 38.5% of patients treated with ECU were administered an estimated dose higher than the FDA-approved label recommends.30
There is a limited amount of real-world evidence available for C5i treatments for PNH, but the potential inadequate therapeutic response of C5i treatments suggested from this study is supported by a phase 3 study of ECU (n = 121) and RAV (n = 126) in the treatment of PNH.13 That study demonstrated that 26.4% of patients treated with RAV and 33.9% of patients treated with ECU required at least 1 transfusion within the 6-month treatment period.13 That same study also indicated that patients with PNH had poor disease control with 32% and 36% of patients treated with RAV and ECU, respectively, not reaching hemoglobin stabilization defined as avoidance of at least a 2g/dL decrease in hemoglobin level from baseline in the absence of transfusion.13 The recent approval of the C3/C3b inhibitor, pegcetacoplan, by the FDA in May 2021, has the potential to address the unmet need of BTH and anemia by targeting extravascular hemolysis, as well intravascular hemolysis, which may result in better control of disease.18 However, pegcetacoplan claims data were not yet available in this analysis and should be included in future studies.
We found that the total cost, including medical benefit and pharmacy benefit costs, of treating members with RAV was 13.9% lower than ECU, although no formal statistical testing was performed. This total cost difference could be largely attributed to drug costs given the 10% lower per vial Wholesale Acquisition Cost list price of RAV. Using the sensitivity analysis among those with a confirmed PNH diagnosis, the total cost of care was found to be similar to the primary analysis ($741,472 for ECU, n = 23; $625,338 for RAV, n = 47) as were other outcomes, including discontinuation, as shown in Table 2, which adds confidence to our findings. Further, the total cost difference should be viewed cautiously because of new-start loading doses and administration in the hospital outpatient setting, which were much more frequent in the RAV cohort, compared with ECU. These 2 differences indicate there is greater potential for a higher total cost of care associated with RAV treatment that may not be captured herein.
Of the 35 members who switched from ECU to RAV during the study period, we found the number of BTH episodes and number of transfusions were similar, indicating that disease control remained similar despite switching treatments. These findings are supported by a phase 3 clinical trial, which investigated members switching from ECU to RAV (n = 97) and those continuing treatment with ECU (n = 98).14 The trial found 12.4% of members who switched to RAV required at least 1 transfusion, compared with 17.3% who remained on ECU.14
LIMITATIONS
This study is subject to some limitations. This study describes treatment patterns with both ECU and RAV; however, there was no attempt to control for confounding or conduct statistical testing on differences between members treated with each therapy. As such, further research is needed to clarify any differences between treatment groups. Classifying treatment naivety using claims data is imperfect. Claims include assumptions of members’ actual drug use and diagnoses. Dosages were calculated using the dose per day average method calculated over the members’ C5i treatment post-index period; therefore, a member’s highest dosage is not independently assessed. Additionally, dose escalation was only determined after members reached maintenance dosing, and members who escalated prior to this period would not be counted. Discontinuation was defined as a gap in therapy of at least twice the approved dosing interval for each product; however, this definition has not been validated in other studies and was based on clinical input. The data used in this study were limited to a commercial population, and results may not be generalizable to Medicare or Medicaid populations. Costs in this analysis are limited to health care claim expenses. However, PNH also results in significant indirect and societal costs. The optimal measure of BTH would include laboratory measures of lactate dehydrogenase, along with clinical symptoms, and may not be fully captured in the present analysis given coding limitations. Additionally, we did not evaluate complement-amplifying conditions (eg, infections) that may explain causes for BTH. The total sample size in the analysis is small; however, this study sample is comparable or larger than in previous PNH studies. Consistent with prior PNH literature, an inclusion by exclusion approach was used to identify members for this analysis and thus the total population may be underestimated.19 Also, 11 of the 34 members in the ECU cohort and 5 of the 52 members in the RAV cohort did not have a medical claim with an ICD-10-CM code for PNH, thus their PNH diagnosis is not definite. That said, sensitivity analyses were performed among those with a definite PNH diagnosis. Because individual weights were not available and RAV dosing is weight-based, standard of care based on dosage had limited applicability for the RAV cohort. Eculizumab requires 5 loading doses and RAV a single loading dose, which increases treatment costs during the induction phase. Therefore, cost comparisons that include members new to therapy should be considered.
Conclusions
In this retrospective descriptive study, the real-world annual total costs for C5i-treated members were more than $625,000 per treated individual. Drug costs accounted for greater than 75% of the total cost of care. C5i discontinuation was observed in about one-third of the sample, and there is a subset of members who do not appear to have their condition adequately controlled as they experienced dose escalation above the FDA-approved dose, experienced BTH, and/or remained transfusion-dependent. With these high PNH drug therapy costs and frequent inadequate clinical outcomes, it is imperative managed care pharmacists remain vigilant in monitoring these members to aid in optimizing cost-effective pharmacotherapy. Further studies are needed to understand the reasons for these observed findings.
ACKNOWLEDGMENTS
The authors thank Nicholas Redman and Rachael Mann, employees of Labcorp Drug Development, Inc., for support in the development of this manuscript.
REFERENCES
- 1.Jalbert JJ, Chaudhari U, Zhang H, Weyne J, Shammo JM. Epidemiology of PNH and real-world treatment patterns following an incident PNH diagnosis in the US. Blood. 2019;134 Suppl 1:3407. doi:10.1182/blood-2019-125867 [Google Scholar]
- 2.Bektas M, Copley-Merriman C, Khan S, Sarda SP, Shammo JM. Paroxysmal nocturnal hemoglobinuria: Role of the complement system, pathogenesis, and pathophysiology. J Manag Care Spec Pharm. 2020;26 Suppl 12-b:S3-8. doi:10.18553/jmcp.2020.26.12-b.s3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hill A, DeZern AE, Kinoshita T, Brodsky RA. Paroxysmal nocturnal haemoglobinuria. Nat Rev Dis Primers. 2017;3:17028. doi:10.1038/nrdp.2017.28 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Horiuchi T, Tsukamoto H. Complement-targeted therapy: Development of C5- and C5a-targeted inhibition. Inflamm Regen. 2016;36(1):11. doi:10.1186/s41232-016-0013-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ricklin D, Hajishengallis G, Yang K, Lambris JD. Complement: A key system for immune surveillance and homeostasis. Nat Immunol. 2010;11(9):785-97. doi:10.1038/ni.1923 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sahin F, Yilmaz AF, Ozkan MC, Gokmen NM, Saydam G. PNH. is a debilitating, fatal but treatable disease: Same disease, different clinical presentations. Am J Blood Res. 2015;5(1):30-3. [PMC free article] [PubMed] [Google Scholar]
- 7.US Food and Drug Administration. Rare diseases at FDA. Accessed May 23, 2023. https://www.fda.gov/patients/rare-diseases-fda
- 8.Mullard A. 2011 FDA drug approvals. Nat Rev Drug Discov. 2012;11(2):91-4. doi:10.1038/nrd3657 [DOI] [PubMed] [Google Scholar]
- 9.Mullard A. 2021 FDA approvals. Nat Rev Drug Discov. 2022;21(2):83-8. doi:10.1038/d41573-022-00001-9 [DOI] [PubMed] [Google Scholar]
- 10.Rome BN, Egilman AC, Kesselheim AS. Trends in prescription drug launch prices, 2008-2021. JAMA. 2022;327(21):2146-7. doi:10.1001/jama.2022.5542 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bektas M, Copley-Merriman C, Khan S, Sarda SP, Shammo JM. Paroxysmal nocturnal hemoglobinuria: Current treatments and unmet needs. J Manag Care Sec Pharm. 2020;26 Suppl 12-b:S14-20. doi:10.18553/jmcp.2020.26.12-b.s14 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Risitano AM, Rotoli B. Paroxysmal nocturnal hemoglobinuria: Pathophysiology, natural history and treatment options in the era of biological agents. Biologics. 2008;2(2):205-22. doi:10.2147/btt.s1420 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lee JW, Sicre de Fontbrune F, Wong Lee Lee L, et al. Ravulizumab (ALXN1210) vs eculizumab in adult patients with PNH naive to complement inhibitors: The 301 study. Blood. 2019;133(6):530-9. doi:10.1182/blood-2018-09-876136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kulasekararaj AG, Hill A, Rottinghaus ST, et al. Ravulizumab (ALXN1210) vs eculizumab in C5-inhibitor–experienced adult patients with PNH: The 302 study. Blood. 2019;133(6):540-9. doi:10.1182/blood-2018-09-876805 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hill AR, Rother RP, Arnold L, et al. Eculizumab prevents intravascular hemolysis in patients with paroxysmal nocturnal hemoglobinuria and unmasks low-level extravascular hemolysis occurring through C3 opsonization. Haematologica. 2010;95(4):567-73. doi:10.3324/haematol.2009.007229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Harder MJ, Kuhn N, Schrezenmeier H, et al. Incomplete inhibition by eculizumab: Mechanistic evidence for residual C5 activity during strong complement activation. Blood. 2017;129(8):970-80. doi:10.1182/blood-2016-08-732800 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Brodsky RA, Peffault de Latour R, Rottinghaus ST, et al. Characterization of breakthrough hemolysis events observed in the phase 3 randomized studies of ravulizumab versus eculizumab in adults with paroxysmal nocturnal hemoglobinuria. Haematologica. 2021;106(1):230-7. doi:10.3324/haematol.2019.236877 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rehan ST, Hashmi MR, Asghar MS, Tahir MJ, Yousaf Z. Pegcetacoplan - Anovel C3 inhibitor for paroxysmal nocturnal hemoglobinuria. Health Sci Rep. 2022;5(3):e512. doi:10.1002/hsr2.512 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cheng WY, Sarda SP, Mody-Patel N, et al. Real-world eculizumab dosing patterns among patients with paroxysmal nocturnal hemoglobinuria in a US population. Clinicoecon Outcomes Res. 2022;14:357-69. doi:10.2147/ceor.S346816 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Cheng WY, Sarda SP, Mody-Patel N, et al. Real-world healthcare resource utilization (HRU) and costs of patients with paroxysmal nocturnal hemoglobinuria (PNH) receiving eculizumab in a US population. Adv Ther. 2021;38(8):4461-79. doi:10.1007/s12325-021-01825-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hill A, Piatek CI, Peffault de Latour R, et al. Breakthrough hemolysis in adult patients with paroxysmal nocturnal hemoglobinuria treated with ravulizumab: Results of a 52-week extension from two phase 3 studies. Blood. 2019;134:952. doi:10.1182/blood-2019-128929 [Google Scholar]
- 22.Griffin M, Munir T. Management of thrombosis in paroxysmal nocturnal hemoglobinuria: A clinician’s guide. Ther Adv Hematol. 2017;8(3):119-26. doi:10.1177/2040620716681748 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cançado RD, da Silva Araújo A, Sandes AF, et al. Consensus statement for diagnosis and treatment of paroxysmal nocturnal haemoglobinuria. Hematol Transfus Cell Ther. 2021;43(3):341-8. doi:10.1016/j.htct.2020.06.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Glasheen WP, Cordier T, Gumpina R, Haugh G, Davis J, Renda A. Charlson comorbidity index: ICD-9 update and ICD-10 translation. Am Health Drug Benefits. 2019;12(4):188-97. [PMC free article] [PubMed] [Google Scholar]
- 25.Fisher MD, Watson C, Fox KM, Chen Y-W, Gandra SR. Dosing patterns of three tumor necrosis factor blockers among patients with rheumatoid arthritis in a large United States managed care population. Curr Med Res Opin. 2013;29(5):561-8. doi:10.1185/03007995.2013.786693 [DOI] [PubMed] [Google Scholar]
- 26.Harrison DJ, Huang X, Globe D. Dosing patterns and costs of tumor necrosis factor inhibitor use for rheumatoid arthritis. Am J Health Syst Pharm. 2010;67(15):1281-7. doi:10.2146/ajhp090487 [DOI] [PubMed] [Google Scholar]
- 27.Waters H, Vanderpoel J, McKenzie S, et al. Stability of infliximab dosing in inflammatory bowel disease: Results from a multicenter US chart review. J Med Econ. 2011;14(4):397-402. doi:10.3111/13696998.2011.583152 [DOI] [PubMed] [Google Scholar]
- 28.Wu E, Chen L, Birnbaum H, Yang E, Cifaldi M. Retrospective claims data analysis of dosage adjustment patterns of TNF antagonists among patients with rheumatoid arthritis. Curr Med Res Opin. 2008;24(8):2229-40. doi:10.1185/03007990802229548 [DOI] [PubMed] [Google Scholar]
- 29.Debureaux P-E, Cacace F, Silva BG, et al. Hematological response to eculizumab in paroxysmal nocturnal hemoglobinuria: Application of a novel classification to identify unmet clinical needs and future clinical goals. Blood. 2019;134 Suppl 1:3517. doi:10.1182/blood-2019-125917 [Google Scholar]
- 30.Burke J, Sahli B, Broderick K, Gleason P. Paroxysmal nocturnal hemoglobinuria real-world effectiveness of C5 inhibitors and cost assessment. 4085-B © Prime Therapeutics LLC 10/21. Poster presented at: Presented at The Academy of Managed Care Pharmacy (AMCP) Nexus Meeting; October 20, 2021 Denver, CO. [Google Scholar]