Abstract
Introduction
Pulmonary arterial hypertension (PAH) is a rare, progressive disease resulting from elevated pulmonary arterial pressure leading to right ventricular failure and death. Optimal adherence and persistence to medical therapy are necessary to improve outcomes. The objective of this study was to characterize adherence and persistence to first-line PAH therapies in patients newly initiating treatment.
Methods
This retrospective cohort study utilized Komodo Research Database claims data. Adults initiating therapy were identified based on ≥ 1 claim for a phosphodiesterase 5 inhibitor (PDE5i) and/or an endothelin receptor antagonist (ERA) from January 1, 2017, to June 30, 2022 (index date), continuous medical and pharmacy health plan enrollment for ≥ 12 months before and including index, ≥ 1 inpatient or ≥ 2 outpatient claims for pulmonary hypertension/PAH, and ≥ 1 claim for right heart catheterization. Adherence was measured by proportion of days covered (PDC); nonadherence was defined as PDC < 80%. Persistence was defined as time from index to treatment discontinuation (gap in therapy > 60 days). Propensity score matching was utilized 1:1:1 across groups.
Results
A total of 9176 patients met the study criteria (6989 PDE5i, 1006 ERA, 1181 dual combination). After matching, each cohort included 714 patients. Median (95% confidence interval) persistence was highest for ERA monotherapy (26.5 [19.0–33.0] months), followed by dual combination therapy (19.8 [16.6–23.4] months) and PDE5i monotherapy (12.9 [10.8–17.4] months)—P = 0.019, dual combination versus ERA; P = 0.026, dual combination versus PDE5i. Nonadherence was highest with dual combination therapy (35.4%), followed by PDE5i monotherapy (17.1%) and ERA monotherapy (11.9%)—P < 0.001, dual combination versus each monotherapy.
Conclusions
Adherence to initial PAH therapy is suboptimal, especially with upfront dual combination therapy. Persistence was highest for ERA monotherapy, followed by dual combination therapy and PDE5i monotherapy. Strategies to improve adherence and persistence are crucial to optimizing outcomes.
Graphical Abstract
Keywords: Adherence, Combination therapy, Endothelin receptor antagonist, Persistence, Phosphodiesterase 5 inhibitor, Pulmonary hypertension
Key Summary Points
| Why carry out this study? |
| Optimal adherence and persistence to pulmonary arterial hypertension (PAH) therapy are necessary to improve outcomes; however, adherence and persistence to PAH therapy remain a significant challenge. |
| This retrospective, observational cohort study assessed real-world adherence and persistence among patients newly initiating therapy for PAH. |
| What was learned from the study? |
| Nonadherence (proportion of days covered (PDC) < 80%) was highest with upfront dual combination therapy (endothelin receptor antagonist [ERA] and phosphodiesterase 5 inhibitor [PDE5i]), followed by PDE5i monotherapy, then ERA monotherapy. |
| However, median persistence was highest for ERA monotherapy (26.5 months), followed by dual combination therapy (19.8 months) and PDE5i monotherapy (12.9 months). |
| Strategies to improve adherence and persistence to PAH therapy are crucial to optimize outcomes. |
Digital Features
This article is published with digital features, including a graphical abstract to facilitate understanding of the article. To view digital features for this article go to https://doi.org/10.6084/m9.figshare.29859233.
Introduction
Pulmonary arterial hypertension (PAH) is a rare, severe form of pulmonary hypertension caused by progressive remodeling of the distal pulmonary arterial vasculature, ultimately leading to right ventricular failure and premature death [1]. In the United States, the prevalence of PAH is estimated at ~50 cases per million people [2]. Diagnosis is often delayed by more than 2 years, on average, due to the nonspecific nature of initial symptoms, such as shortness of breath and fatigue. This leads to advanced disease at diagnosis and at time of treatment initiation [3].
The European Society of Cardiology/European Respiratory Society 2022 guidelines for pulmonary hypertension (PH) recommend treatment for PAH based on a patient’s 1-year risk of mortality [4]. Upfront dual combination therapy with a phosphodiesterase 5 inhibitor (PDE5i) and an endothelin receptor antagonist (ERA) is recommended for low- and intermediate-risk patients without cardiopulmonary comorbidities. The 2024 World Symposium on Pulmonary Hypertension (WSPH) consensus statement also recommends upfront dual combination therapy in these patients [5].
Adherence and persistence to PAH therapy remain a significant challenge, however. Two recent meta-analyses reveal wide-ranging variability in adherence. A meta-analysis of 14 studies (n = 14,861) reported that 60.9% (95% confidence interval [CI] 52.3–69.1) of patients were adherent while 42.3% (95% CI 31.6–53.3) discontinued their PAH therapy [6]. In another meta-analysis (20 studies; n = 22,675), the proportion of patients with high medication adherence ranged from 40% (95% CI 35–45) to 94% (95% CI 88–97) [7]. Among the few US studies on this topic are two retrospective claims analyses [8, 9]. In their analysis of the IQVIA Adjudicated Health Plan Database, Frantz et al. reported that the adherence rate was higher for patients on ERA therapy (75.6%; 571/755 patients) than for those on a PDE5i (61.5%; 970/1578). They also found that in both groups, hospitalizations declined as adherence improved [8]. Studer et al. reported similar findings in their analysis of the Optum Research Database (n = 1878), with 76.1% of patients being adherent [9]. In addition, ~40% of patients discontinued their index therapy, with 80.3% of these discontinuing within the first year [9]. However, both of these studies were published in 2020, and treatment guidelines have evolved. Therefore, there is a need for more recent data to assess current trends in PAH medication adherence and persistence and better inform strategies that address changing patient needs.
The issue of PAH medication adherence and persistence is gaining support and visibility. At the 2024 WSPH, adherence was discussed by the special populations task force [10]. The task force recommended patient-centered approaches to adherence, including continuous communication between patients and healthcare providers and patient education, both of which have been effective in increasing persistence [11]. It also recommended routine assessment of medication adherence in PH clinics and its integration into patient-reported outcomes and quality-of-life measures [10]. Thus, evaluating real-world adherence and persistence rates among patients initiating therapy for PAH is critical for identifying those at greatest risk for suboptimal outcomes and for informing the development of targeted support interventions. This study aims to address this gap by analyzing contemporary adherence and persistence rates in patients newly initiating therapy for PAH.
Methods
Study Design and Data Source
This was a retrospective, observational cohort study using closed medical and pharmacy claims data from the Komodo Research Database (KRD), from January 1, 2016, to June 30, 2023 (study period). The KRD includes data from ~330 million patients across the United States spanning from 2015 to the present day, providing US census-level representativeness. The data include inpatient and outpatient medical and pharmacy claims, as well as plan enrollment information.
Study Population
Patients initiating therapy with a PDE5i (sildenafil or tadalafil) and/or an ERA (macitentan or ambrisentan) between January 1, 2017, and June 30, 2022, were identified. Those newly initiating therapy were identified by a 12-month washout period (i.e., no pharmacy claims for PAH therapy in the past 12 months) prior to first therapy. Additionally, patients were required to have continuous medical and pharmacy health plan enrollment for ≥ 1 year prior to and including the index date (baseline period) and ≥ 30 days after the index date, and at least one PH/PAH diagnosis based on ≥ 1 inpatient or ≥ 2 outpatient claims (International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] codes I27.0, I27.20, I27.21, I27.89) on two separate days in the 10 months before first therapy. Patients also had to be ≥ 18 years of age. Exclusion criteria included any claim for a PAH medication during the baseline period or for any PAH medication other than a PDE5i or ERA on the index date, and diagnosis/procedures for chronic thromboembolic pulmonary hypertension, pulmonary endarterectomy, balloon pulmonary angioplasty, lung transplant, or atrial septostomy during the baseline period. PAH-approved PDE5i doses were required; patients with an average of < 1 PDE5i tablet per day between the index date and day 30 were excluded. Those who had not undergone a right heart catheterization within the baseline or follow-up period were also not eligible for the study. Follow-up extended from the index date to the earliest of health plan disenrollment, death (as estimated from available data), or study end.
Three cohorts were created: dual combination therapy, ERA monotherapy, and PDE5i monotherapy. The dual combination therapy cohort included patients with claims for both a PDE5i and an ERA, with the second agent added within 30 days after the index agent. The PDE5i and ERA monotherapy cohorts included patients who received respective PDE5i or ERA monotherapy at index, with no addition of a second agent within 30 days after index.
Variables
Baseline demographic and clinical characteristics included age, sex, race/ethnicity, geographic region, insurance type, and Charlson Comorbidity Index score.
Adherence was measured as the proportion of days covered (PDC), with nonadherence defined as PDC < 80%. PDC was assessed at three time points: (1) during the persistence period (from the index date to treatment discontinuation, as defined below) (2) at 6 months post-index, and (3) at 12 months post-index. PDC was calculated as the number of days with medication available (numerator) divided by the total number of days in the observation period (denominator), multiplied by 100 to express the result as a percentage. For both the monotherapy and dual combination cohorts, the denominator was defined as the number of days from the index date to treatment discontinuation, 183 days, or 365 days, depending on the time point. For the monotherapy cohorts, the numerator was the number of days of medication coverage with either a PDE5i or ERA. For the dual combination cohort, the numerator was the number of days with coverage for both a PDE5i and an ERA.
Persistence was defined as the time from index to treatment discontinuation. For the monotherapy cohorts, treatment discontinuation was considered to be a gap of > 60 days in PDE5i supply or in ERA supply. For the dual combination cohort, treatment discontinuation was considered to be a gap of > 60 days in either PDE5i or ERA supply. The treatment period was determined using dispensing dates and days’ supply of PDE5i and ERA medications from pharmacy claims, with early refills of the same medication appended to start on the day after the end of the previous fill’s supply. Treatment episodes allowed within-class drug switches. Therapy was presumed to continue during hospitalizations.
Statistical Analysis
Demographics, baseline clinical characteristics in the year prior to the index date, and adherence were summarized using descriptive statistics. PDC was analyzed as both a continuous and categorical variable. Baseline comparisons among the three cohorts were performed with P values from Welch’s two sample t-test, Pearson’s chi-squared test, or Fisher’s exact test, depending on data type (e.g., continuous or categorical).
Propensity score matching (PSM) was used to ensure comparability between patients in the monotherapy and combination therapy cohorts in a 1:1:1 ratio. The matching was conducted using the nearest neighbor algorithm without replacement. After matching, group balance was assessed and reported as P values.
Kaplan–Meier analysis was performed to assess time to nonadherence (i.e., PDC < 80%) and time to nonpersistence (measured as the index date to the last prescription date’s end of days’ supply plus 60 days to ensure no further refills), with statistically significant differences between the dual combination and monotherapy cohorts assessed by log-rank test. As a sensitivity analysis, adherence was assessed at 6 and 12 months for patients who had a minimum of 6 and 12 months’ follow-up, respectively, and for all patients.
Statistical analyses were conducted in R 4.1.3 with the packages AdhereR 0.8.1 and survival 3.5–5.
Ethics
As the study was retrospective and used only deidentified aggregated patient data, it was exempt from ethical or institutional review board review/approval. The Komodo Health Research Database is compliant with the Health Insurance Portability and Accountability Act of 1996. The data were used under license from Komodo.
Results
Patient Characteristics
A total of 9176 patients met the study criteria (Fig. 1): 1181 in the dual combination therapy cohort, 6989 in the PDE5i monotherapy cohort, and 1006 in the ERA monotherapy cohort.
Fig. 1.
Patient sample selection. CTEPH chronic thromboembolic pulmonary hypertension, ERA endothelin receptor antagonist, PAH pulmonary arterial hypertension, PDE5i phosphodiesterase 5 inhibitor; PH pulmonary hypertension
Baseline characteristics are shown in Table 1 (before PSM) and Table 2 (after PSM). Age and sex distribution were typical for patients with PAH. After PSM, each cohort included 714 patients. Among the matched cohorts, mean age ranged from 57.5 to 58.9 years and female sex from 67.8 to 75.1%. The comorbidity burden was high in all cohorts. Mean Charlson Comorbidity Index score ranged from 4.1 to 4.6, and between 26.2 and 33.6% of each cohort had a score of ≥ 6. Median (interquartile range) follow-up was 24.6 (13.2–41.1) months in the dual combination therapy cohort, 22.1 (12.6–37.0) months in the PDE5i monotherapy cohort, and 22.3 (11.3–39.2) months in the ERA monotherapy cohort.
Table 1.
Baseline demographic and clinical characteristics (unmatched cohorts)
| Dual combination therapy (n = 1181) | PDE5i monotherapy (n = 6989) | ERA monotherapy (n = 1006) | P valuea | P valueb | |
|---|---|---|---|---|---|
| Age, mean (SD), years | 55.0 (14.0) | 60.9 (14.1) | 58.8 (13.8) | < 0.001 | < 0.001 |
| Female sex, n (%) | 854 (72.3) | 4295 (61.5) | 727 (72.3) | < 0.001 | 0.3 |
| Race/ethnicity, n (%) | < 0.001 | 0.009 | |||
| Asian or Pacific Islander | 48 (4.1) | 212 (3.0) | 34 (3.4) | ||
| Black/African American | 198 (16.8) | 1673 (23.9) | 230 (22.9) | ||
| Hispanic/Latino | 169 (14.3) | 870 (12.4) | 132 (13.1) | ||
| Other | 58 (4.9) | 234 (3.3) | 36 (3.6) | ||
| White | 596 (50.5) | 3440 (49.2) | 472 (46.9) | ||
| Geographic region, n (%) | < 0.001 | < 0.001 | |||
| Northeast | 252 (21.3) | 1555 (22.2) | 180 (17.9) | ||
| South | 319 (27.0) | 2349 (33.6) | 415 (41.3) | ||
| Midwest | 264 (22.4) | 1542 (22.1) | 161 (16.0) | ||
| West | 343 (29.0) | 1496 (21.4) | 232 (23.1) | ||
| Puerto Rico | 1 (0.1) | 42 (0.6) | 15 (1.5) | ||
| Insurance type, n (%) | < 0.001 | < 0.001 | |||
| Commercial | 417 (35.3) | 1935 (27.7) | 330 (32.8) | ||
| Medicaid | 437 (37.0) | 1802 (25.8) | 265 (26.3) | ||
| Medicare | 299 (25.3) | 3077 (44.0) | 394 (39.2) | ||
| Medicare medical, commercial drug | 25 (2.1) | 162 (2.3) | 14 (1.4) | ||
| Index year, n (%) | 0.027 | 0.011 | |||
| 2017 | 153 (13.0) | 750 (10.7) | 145 (14.4) | ||
| 2018 | 193 (16.3) | 1286 (18.4) | 187 (18.6) | ||
| 2019 | 229 (19.4) | 1479 (21.2) | 228 (22.7) | ||
| 2020 | 204 (17.3) | 1246 (17.8) | 157 (15.6) | ||
| 2021 | 252 (21.3) | 1486 (21.3) | 202 (20.1) | ||
| 2022 | 150 (12.7) | 742 (10.6) | 87 (8.6) | ||
| CCI | < 0.001 | 0.3 | |||
| Mean (SD) | 4.1 (2.8) | 4.9 (3.0) | 4.2 (2.9) | ||
| CCI, n (%) | < 0.001 | 0.33 | |||
| 0–1 | 160 (13.5) | 607 (8.7) | 136 (13.5) | ||
| 2–3 | 468 (39.6) | 2217 (31.7) | 391 (38.9) | ||
| 4–5 | 262 (22.2) | 1522 (21.8) | 200 (19.9) | ||
| ≥ 6 | 291 (24.6) | 2643 (37.8) | 279 (27.7) | ||
| PDE5i use in first 30 days, n (%) | < 0.001 | < 0.001 | |||
| Sildenafil | 498 (42.2) | 5668 (81.1) | 0 | ||
| Tadalafil | 665 (56.3) | 1281 (18.3) | 0 | ||
| Sildenafil + tadalafil | 18 (1.5) | 40 (0.6) | 0 | ||
| ERA use in first 30 days, n (%) | < 0.001 | < 0.001 | |||
| Ambrisentan | 731 (61.9) | 0 | 439 (43.6) | ||
| Bosentan | 8 (0.7) | 0 | 31 (3.1) | ||
| Macitentan | 437 (37.0) | 0 | 535 (53.2) | ||
| Ambrisentan + bosentan | 0 | 0 | 1 (0.1) | ||
| Ambrisentan + macitentan | 5 (0.4) | 0 | 0 | ||
| Daily pill burden at index date | < 0.001 | 0.016 | |||
| Mean (SD) | 9.7 (6.4) | 13.0 (8.5) | 9.0 (7.0) | ||
| Median (IQR) | 8.5 (5.0–13.0) | 11.3 (7.0–17.0) | 7.8 (4.0–12.5) |
CCI Charlson Comorbidity Index, ERA endothelin receptor antagonist, IQR interquartile range, PDE5i phosphodiesterase 5 inhibitor, SD standard deviation
aComparison of dual combination therapy and PDE5i monotherapy cohorts
bComparison of dual combination therapy and ERA monotherapy cohorts
Table 2.
Baseline demographic and clinical characteristics (matched cohorts)
| Dual combination therapy (n = 714) | PDE5i monotherapy (n = 714) | ERA monotherapy (n = 714) | P valuea | P valueb | |
|---|---|---|---|---|---|
| Age, mean (SD), years | 57.5 (14.0) | 58.9 (13.5) | 58.1 (13.8) | 0.055 | 0.41 |
| Female sex, n (%) | 536 (75.1) | 484 (67.8) | 530 (74.2) | 0.003 | 0.76 |
| Race/ethnicity, n (%) | > 0.99 | 0.88 | |||
| Asian or Pacific Islander | 27 (3.8) | 24 (3.4) | 29 (4.1) | ||
| Black/African American | 156 (21.8) | 159 (22.3) | 160 (22.4) | ||
| Hispanic/Latino | 102 (14.3) | 102 (14.3) | 107 (15.0) | ||
| Other | 37 (5.2) | 39 (5.5) | 29 (4.1) | ||
| White | 392 (54.9) | 390 (54.6) | 389 (54.5) | ||
| Geographic region, n (%) | 0.11 | 0.61 | |||
| Northeast | 132 (18.5) | 161 (22.5) | 130 (18.2) | ||
| South | 245 (34.3) | 207 (29.0) | 266 (37.3) | ||
| Midwest | 141 (19.7) | 151 (21.1) | 127 (17.8) | ||
| West | 195 (27.3) | 192 (26.9) | 188 (26.3) | ||
| Puerto Rico | 1 (0.1) | 3 (0.4) | 3 (0.4) | ||
| Insurance type, n (%) | 0.69 | 0.82 | |||
| Commercial | 225 (31.5) | 206 (28.9) | 221 (31.0) | ||
| Medicaid | 232 (32.5) | 243 (34.0) | 219 (30.7) | ||
| Medicare | 245 (34.3) | 250 (35.0) | 261 (36.6) | ||
| Medicare medical, commercial drug | 12 (1.7) | 15 (2.1) | 13 (1.8) | ||
| Index year, n (%) | 0.81 | > 0.99 | |||
| 2017 | 92 (12.9) | 91 (12.7) | 93 (13.0) | ||
| 2018 | 122 (17.1) | 127 (17.8) | 118 (16.5) | ||
| 2019 | 154 (21.6) | 143 (20.0) | 157 (22.0) | ||
| 2020 | 118 (16.5) | 124 (17.4) | 121 (16.9) | ||
| 2021 | 161 (22.5) | 149 (20.9) | 160 (22.4) | ||
| 2022 | 67 (9.4) | 80 (11.2) | 65 (9.1) | ||
| CCI | 0.014 | 0.62 | |||
| Mean (SD) | 4.2 (2.9) | 4.6 (2.9) | 4.1 (2.8) | ||
| CCI, n (%) | 0.023 | 0.79 | |||
| 0–1 | 93 (13.0) | 85 (11.9) | 104 (14.6) | ||
| 2–3 | 274 (38.4) | 240 (33.6) | 271 (38.0) | ||
| 4–5 | 160 (22.4) | 149 (20.9) | 149 (20.9) | ||
| ≥ 6 | 187 (26.2) | 240 (33.6) | 190 (26.6) | ||
| PDE5i use in first 30 days, n (%) | < 0.001 | < 0.001 | |||
| Sildenafil | 299 (41.9) | 566 (79.3) | 0 | ||
| Tadalafil | 403 (56.4) | 144 (20.2) | 0 | ||
| Sildenafil + tadalafil | 12 (1.7) | 4 (0.6) | 0 | ||
| ERA use in first 30 days, n (%) | < 0.001 | < 0.001 | |||
| Ambrisentan | 411 (57.6) | 0 | 317 (44.4) | ||
| Bosentan | 6 (0.8) | 0 | 22 (3.1) | ||
| Macitentan | 294 (41.2) | 0 | 374 (52.4) | ||
| Ambrisentan + bosentan | 0 | 0 | 1 (0.1) | ||
| Ambrisentan + macitentan | 3 (0.4) | 0 | 0 | ||
| Daily pill burden at index date | < 0.001 | < 0.001 | |||
| Mean (SD) | 10.3 (6.5) | 12.4 (7.9) | 9.0 (7.0) | ||
| Median (IQR) | 9.0 (5.0–13.9) | 11.0 (6.3–16.3) | 7.9 (4.0–12.5) |
CCI Charlson Comorbidity Index, ERA endothelin receptor antagonist, IQR interquartile range, PDE5i phosphodiesterase 5 inhibitor, SD standard deviation
aComparison of dual combination therapy and PDE5i monotherapy cohorts
bComparison of dual combination therapy and ERA monotherapy cohorts
Adherence
Overall adherence in terms of mean PDC was 80.6%, 90.0%, and 91.9% for the dual combination therapy, PDE5i monotherapy, and ERA monotherapy cohorts, respectively. There was a statistically significant difference between the dual combination therapy cohort and each of the monotherapy cohorts (P < 0.001, both comparisons), favoring monotherapy. A substantial proportion of patients were nonadherent (PDC < 80%) during the study period: 35.4% in the dual combination therapy cohort, 17.1% in the PDE5i monotherapy cohort, and 11.9% in the ERA monotherapy cohort (P < 0.001, comparison of dual combination therapy to each monotherapy). Median time to PDC < 80% was statistically significantly shorter for the dual combination therapy cohort (0.03 months [95% CI 0.03–0.03]) compared with the PDE5i monotherapy (3.71 months [95% CI 3.71–4.96]) and ERA monotherapy (8.64 months [95% CI 6.18–11.10]) cohorts (P < 0.001, both comparisons; Fig. 2).
Fig. 2.
Kaplan–Meier analysis of time to PDC < 80% for dual combination therapy compared with PDE5i monotherapy and ERA monotherapy. ERA endothelin receptor antagonist, PDC proportion of days covered, PDE5i, phosphodiesterase 5 inhibitor
While adherence in terms of mean PDC was ≥ 80% for all cohorts over the entire persistence period, mean PDC was < 80% for all cohorts at both 6 months and 12 months of follow-up. Among patients with a minimum of 6 months’ follow-up, the proportion adherent at 6 months was 67.8%, 71.3%, and 78.1% for the dual combination therapy, PDE5i monotherapy, and ERA monotherapy cohorts, respectively. There was a statistically significant difference between the dual combination therapy cohort and each of the monotherapy cohorts (P < 0.001, both comparisons; Fig. 3a and b). Among patients with a minimum of 12 months’ follow-up, the proportion of patients adherent at 12 months was 65.0%, 65.9%, and 73.1% for dual combination therapy, PDE5i monotherapy, and ERA monotherapy cohorts, respectively, with a statistically significant difference between the dual combination therapy cohort and the ERA monotherapy cohort (P < 0.001; Fig. 3a and b).
Fig. 3.
Mean PDC overall, at 6 months (among patients with a minimum of 6 months of follow-up and all patients), and at 12 months (among patients with a minimum of 12 months of follow-up and all patients) for dual combination therapy compared with (a) PDE5i monotherapy and (b) ERA monotherapy. ERA endothelin receptor antagonist, PDC proportion of days covered, PDE5i phosphodiesterase 5 inhibitor
Persistence
Median persistence was 19.8 months (95% CI 16.6–23.4) in the dual combination therapy cohort, 12.9 months (95% CI 10.8–17.4) in the PDE5i monotherapy cohort, and 26.5 months (95% CI 19.0–33.0) in the ERA monotherapy cohort (Fig. 4). There was a statistically significant difference in persistence between the dual combination therapy cohort and both the PDE5i monotherapy cohort (P = 0.026; Fig. 4a) and the ERA monotherapy cohort (P = 0.019; Fig. 4b). The proportion of patients who discontinued their medication during study follow-up was 52.9%, 55.6%, and 45.0% in the dual combination therapy, PDE5i monotherapy, and ERA monotherapy cohorts, respectively.
Fig. 4.
Kaplan–Meier analysis of time to discontinuation for dual combination therapy compared with (a) PDE5i monotherapy and (b) ERA monotherapy. ERA endothelin receptor antagonist, PDE5i phosphodiesterase 5 inhibitor
Discussion
This study examined medication adherence and persistence in 2142 patients with PAH newly initiating upfront dual combination therapy (PDE5i and ERA), PDE5i monotherapy, or ERA monotherapy between 2017 and 2022. Although overall adherence was high in each cohort (80.6%, 90.0%, and 91.9%, respectively), a substantial proportion of patients were nonadherent (PDC < 80%) over the study period (35.4%, 17.1%, and 11.9%, respectively). Persistence was highest for ERA monotherapy (26.5 months), followed by dual combination therapy (19.8 months) and PDE5i monotherapy (12.9 months). Consistent with the persistence to therapy, the proportion of patients who discontinued their medication during study follow-up was lowest in the ERA monotherapy cohort (45.0%), followed by dual combination therapy (52.9%) and PDE5i monotherapy (55.6%).
Persistence findings in this study align with previously reported results showing that ~40% of patients discontinued their PAH therapy [6, 9]. Studer et al. found that patients on combination therapy were less likely to discontinue therapy than those on monotherapy [9]. However, the study did not use matched cohorts, so the proportion of patients on combination therapy (9.2%) was much lower than the proportion on monotherapy (90.8%).
Mean PDC rates in the current study are higher than those previously reported from other retrospective US claims database studies, which have shown mean PDC for PAH therapies ranging from 60 to 80% [9, 12, 13]. Similarly, the proportions of patients considered to be nonadherent (i.e., PDC < 80%) in the present study (11.9–35.4%) are considerably lower than in prior studies (24–53%) [8, 13]. This is especially important given the recent study by De Marco et al. (under review), who found that the odds of morbidity and mortality are two to three times higher when patients are nonadherent to PDE5i and ERA therapy.
Persistence represents a complementary yet distinct dimension of medication-taking behavior. While adherence captures whether patients take their medications as prescribed on a daily basis, persistence measures whether they continue with the therapy over time. In the current study, approximately half of all patients discontinued their PAH therapy during the follow-up period, suggesting that many patients are discontinuing treatment within the first year. These discontinuations may reflect adverse events, perceived inefficacy, cost, or system-level barriers, all of which merit further investigation.
An important contributor to both poor medication adherence and early discontinuation is the high overall pill burden facing patients with PAH. In this study, patients were taking an average of ~10 tablets per day at index, including medications for PAH as well as for comorbid conditions. This cumulative pill burden may lead to unintentional nonadherence (e.g., missed doses due to timing confusion) or intentional nonadherence (e.g., stopping medication due to fatigue or perceived futility). It may also increase the likelihood of early discontinuation resulting from side effects or medication interactions. Reducing regimen complexity, through fixed-dose combinations or once-daily medications, may be an effective strategy to support both adherence and persistence.
It is worth noting that, even after PSM, patients in the PDE5i monotherapy cohort had the highest overall daily pill burden (mean 12.4 pills), followed by the dual therapy cohort (10.3 pills), with the ERA cohort having the lowest overall burden (9.0 pills). The higher pill burden is likely a result of most (79%) patients in the PDE5i monotherapy cohort receiving sildenafil, which is usually taken three times daily, while more than half (56%) of the patients in the dual combination therapy cohort received tadalafil, which is taken once daily. This difference in pill burden and dosing frequency may have influenced treatment behavior in complex ways: although adherence was higher in the PDE5i monotherapy cohort than in the dual therapy cohort, persistence was lower. It is possible that the higher-frequency dosing of sildenafil supported short-term adherence by reinforcing routine use but became more burdensome over time, contributing to earlier discontinuation. Conversely, patients on dual therapy may have had lower adherence due to regimen complexity but may have persisted longer due to greater perceived therapeutic benefit or once-daily dosing of tadalafil.
PDC is preferred over alternative metrics, such as the medication possession ratio, because it accounts for overlapping days supplied from early refills and is capped at 100% [14, 15]. This approach offers a more conservative and standardized measure of adherence than the medication possession ratio, facilitating comparisons across studies. Unlike the medication possession ratio, PDC distinguishes between adherence and persistence by excluding periods following treatment discontinuation from the adherence calculation. Therefore, metrics of adherence and persistence should be evaluated jointly to fully characterize treatment behavior.
The threshold of PDC ≥ 80% has been widely adopted as a standard measure of adequate medication adherence and originated from early empirical research in the management of hypertension [16, 17]. For once-daily medications, this corresponds to permitting up to six missed doses per month while still being considered adherent.
In the context of PAH—a progressive, fatal disease—this threshold may not ensure adequate drug exposure to achieve optimal clinical outcomes. Moreover, the clinical impact of missing doses may depend not only on the total number of missed doses, but also on whether they are missed intermittently or consecutively. This distinction may be especially relevant given the pharmacokinetic profiles of PAH therapies. For example, the ERA bosentan and the PDE5i sildenafil have relatively short half-lives (~4 to 5 h) and are typically taken twice daily and three times daily, respectively [18, 19]. In contrast, the once-daily PDE5i tadalafil has a longer half-life of ~35 h in patients with PAH who are not taking concomitant bosentan [20]. Among once-daily ERAs, ambrisentan has a half-life of ~15 h, while macitentan has a parent drug half-life of ~16 h and an active metabolite with a substantially longer half-life of ~48 h [21, 22].
Importantly, while PDC considers whether patients have enough medication to take all prescribed doses on a given day, it does not verify that the doses were actually taken or whether they were taken at the appropriate times. Therefore, even with a PDC ≥ 80%, patients on complex regimens may still experience variable drug exposure if doses are skipped, particularly with therapies that have shorter half-lives and are taken more than once daily (e.g., sildenafil or bosentan) [18, 19].
To our knowledge, no studies have made a direct link between PDC-defined adherence thresholds and clinical outcomes in patients with PAH. Additional research is needed to determine whether the conventional ≥ 80% threshold is sufficient or whether a more stringent adherence cutoff is justified in this high-risk population.
The 2024 WSPH recognized lack of adherence and persistence to PAH medication as an important cause of treatment failure and identified significant knowledge gaps in the area [10]. Future research should investigate the underlying reasons and patient mindsets that contribute to nonadherence and poor persistence. As already discussed, the high pill burden facing patients with PAH might contribute to poor adherence and persistence. Simplified regimens and tailored interventions, such as single-tablet combination therapies, may help enhance both adherence and persistence. Patient-centered approaches, including education programs, might also improve adherence and persistence; recent qualitative research suggests that persistence in patients with PAH can be improved through continuous patient engagement, effective communication between healthcare providers and patients, and proactive management of side effects [11]. Consistent with the recommendations of the WSPH task force, regular assessment of medication adherence is essential for prompt identification of nonadherence [10]. As suggested by the task force, incorporating adherence evaluations into patient-reported outcomes and quality-of-life measures could improve continuous monitoring, enabling timely interventions that effectively reduce nonadherence.
Strengths and Limitations
This study has several strengths. For example, it includes a large sample size and utilizes recent data (2016–2023), providing contemporary insights into adherence trends in PAH therapy. However, certain limitations should be acknowledged. As with all studies of this type, the use of claims data introduces potential sources of imprecision, including misclassification of study variables and missing data, which could affect the validity and generalizability of the findings. For example, there are currently no PAH-specific ICD-10-CM codes, and coding errors or incomplete data may result in the underrepresentation of comorbidities. Additionally, adherence was estimated from claims data, but medication dispensing records do not confirm medication consumption. Nevertheless, previous real-world studies have successfully used dispensing data to estimate medication adherence in patients with PAH [12, 23, 24], and claims data are considered the most robust real-world evidence for measuring adherence, unlike patient self-reporting. Finally, as the dataset is limited to insured patients, our findings may not be generalizable to other populations. Despite these limitations, the study provides important real-world evidence on adherence patterns and their implications.
Conclusions
Nonadherence to upfront therapies in PAH can be detrimental to patient outcomes. In this study, nonadherence increased when patients were started on dual combination therapy; this is especially important to note for a rare, fatal, progressive disease in which treatment complexity increases with disease progression. Persistence was highest for the ERA monotherapy cohort, which was in line with the results for adherence, but interestingly, we found dual combination therapy to have the next highest persistence and PDE5i monotherapy to have the lowest persistence. Simplifying regimens in addition to addressing outstanding financial constraints and access barriers is paramount to improving adherence and persistence for these patients. Healthcare providers should focus on shared decision-making between providers and patients as well as actively monitoring patient adherence at each visit.
Acknowledgments
Medical Writing, Editorial, and Other Assistance
Medical writing support was provided by Melanie Jones, BSc, and Allison Michaelis, PhD, on behalf of Twist Medical and was funded by Johnson & Johnson.
Author Contributions
Teresa De Marco and Harrison W. Farber contributed to data analysis or interpretation. Carly J. Paoli contributed to study design and data analysis or interpretation. Nicole S. Croteau and Fei Tang contributed to study design, data collection, and data analysis or interpretation. All authors reviewed and critically revised the manuscript, approved the final draft, and agree to be accountable for the accuracy and integrity of the work.
Funding
Sponsorship for this study and all publication charges were funded by Johnson & Johnson, Titusville, NJ, USA.
Data Availability
The data used in this study were obtained from the Komodo Health Research Database, a proprietary claims and healthcare encounter dataset. Due to data use agreements and licensing restrictions, we are not permitted to share the raw data publicly. Researchers interested in accessing these data must obtain a license directly from Komodo Health (https://www.komodohealth.com/).
Declarations
Conflict of Interest
Teresa De Marco has received research grants from CareDx and Acceleron; has received consultancy fees from Aerovate, Boston Scientific, Atara, Keros, Merck, Natera, Pulnovo, Tectonic, and United Therapeutics; has participated on advisory boards or data safety monitoring boards for Kamada, Merck, Keros Therapeutics, Tectonic, and BIAL; has received honoraria for a Simply Speaking PAH CME lecture; has received support for meeting attendance from CareDx, Atara, Kamada, and United Therapeutics; and has served on an endpoint adjudication committee for Johnson & Johnson. Carly J. Paoli is an employee and stockholder of Johnson & Johnson. Nicole S. Croteau and Fei Tang were employees of Cytel at the time of this study. Cytel has received consultancy fees from Johnson & Johnson for the conduct of this study. Harrison W. Farber has received speaking honoraria from Bayer and scientific advisory board fees from Acceleron (Merck), Actelion (Janssen), Altavant, Aerami, Aerovate, Pulmovant, and United Therapeutics, and is the Adjudicator for Keros.
Ethical Approval
As the study was retrospective and used only deidentified aggregated patient data, it was exempt from ethical or institutional review board review/approval. The Komodo Health Research Database is compliant with the Health Insurance Portability and Accountability Act of 1996. The data were used under license from Komodo.
References
- 1.Ruopp NF, Cockrill BA. Diagnosis and treatment of pulmonary arterial hypertension: a review. JAMA. 2022;327(14):1379–91 (Erratum in: JAMA. 2022;328(9):892). [DOI] [PubMed]
- 2.Leber L, Beaudet A, Muller A. Epidemiology of pulmonary arterial hypertension and chronic thromboembolic pulmonary hypertension: identification of the most accurate estimates from a systematic literature review. Pulm Circ. 2021;11(1): 2045894020977300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Brown LM, Chen H, Halpern S, et al. Delay in recognition of pulmonary arterial hypertension: factors identified from the REVEAL Registry. Chest. 2011;140(1):19–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Humbert M, Kovacs G, Hoeper MM, ESC/ERS Scientific Document Group, et al. 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Respir J. 2023;61(1): 2200879. [DOI] [PubMed] [Google Scholar]
- 5.Chin KM, Gaine SP, Gerges C, et al. Treatment algorithm for pulmonary arterial hypertension. Eur Respir J. 2024;64(4):2401325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Qadus S, Naser AY, Ofori-Asenso R, Ademi Z, Al Awawdeh S, Liew D. Adherence and discontinuation of disease-specific therapies for pulmonary arterial hypertension: a systematic review and meta-analysis. Am J Cardiovasc Drugs. 2023;23(1):19–33. [DOI] [PubMed] [Google Scholar]
- 7.Le Bozec A, Korb-Savoldelli V, Boiteau C, et al. Medication adherence, related factors and outcomes among patients with pulmonary arterial hypertension or chronic thromboembolic pulmonary hypertension: a systematic review. Eur Respir Rev. 2024;33(173): 240006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Frantz RP, Hill JW, Lickert CA, et al. Medication adherence, hospitalization, and healthcare resource utilization and costs in patients with pulmonary arterial hypertension treated with endothelin receptor antagonists or phosphodiesterase type-5 inhibitors. Pulm Circ. 2020;10(1): 2045894019880086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Studer S, Hull M, Pruett J, Elliott C, Tsang Y, Drake W. Retrospective database analysis of treatment patterns among patients with pulmonary arterial hypertension. Pulm Ther. 2020;6(1):79–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Preston IR, Howard LS, Langleben D, et al. Management of pulmonary hypertension in special conditions. Eur Respir J. 2024;64(4):2401180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Shlobin OA, Bruce G, Gomez Rendon G, et al. Practices affEcting macitentan and selexipag patient persistence Rates utilizing pulmonary arterial hypertension clinical Site and patIent perSpecTives (PERSIST): a US qualitative analysis. Pulm Circ. 2024;14(4): e12441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dean BB, Saundankar V, Stafkey-Mailey D, et al. Medication adherence and healthcare costs among patients with pulmonary arterial hypertension treated with oral prostacyclins: a retrospective cohort study. Drugs Real World Outcomes. 2020;7(3):229–39 (Erratum in: Drugs Real World Outcomes. 2020;7(3):241–2). [DOI] [PMC free article] [PubMed]
- 13.Ruiz G, Yeaw J, Lickert CA, et al. Using real world evidence to describe pulmonary arterial hypertension treatment patterns, healthcare resource utilization, and costs associated with PDE-5 inhibitor monotherapy. J Health Econ Outcomes Res. 2018;5(2):206–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Raebel MA, Schmittdiel J, Karter AJ, Konieczny JL, Steiner JF. Standardizing terminology and definitions of medication adherence and persistence in research employing electronic databases. Med Care. 2013;51(8 Suppl 3):S11-21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Canfield SL, Zuckerman A, Anguiano RH, et al. Navigating the wild west of medication adherence reporting in specialty pharmacy. J Manag Care Spec Pharm. 2019;25(10):1073–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Baumgartner PC, Haynes RB, Hersberger KE, Arnet I. A systematic review of medication adherence thresholds dependent of clinical outcomes. Front Pharmacol. 2018;9:1290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gellad WF, Thorpe CT, Steiner JF, Voils CI. The myths of medication adherence. Pharmacoepidemiol Drug Saf. 2017;26(12):1437–41. [DOI] [PubMed] [Google Scholar]
- 18.Tracleer® (bosentan) tablets, for oral use. US prescribing information. Johnson & Johnson; 2024.
- 19.Revatio® (sildenafil) tablets, for oral use. US prescribing information. Viatris Specialty LLC; 2024.
- 20.Adcirca® (tadalafil) tablets, for oral use. US prescribing information. Eli Lilly and Company; 2020.
- 21.Letairis® (ambrisentan) tablets, for oral use. US prescribing information. Gilead Sciences, Inc.; 2025.
- 22.Opsumit® (macitentan) tablets, for oral use. US prescribing information. Johnson & Johnson; 2025.
- 23.Schikowski EM, Swabe G, Chan SY, Magnani JW. Association between copayment and adherence to medications for pulmonary arterial hypertension. J Am Heart Assoc. 2022;11(22): e026620. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Tsai CY, Shen CW, Lai HL, Chen CY. Adherence and treatment patterns of disease-specific drugs among patients with pulmonary arterial hypertension: a nationwide, new-user cohort study. Front Pharmacol. 2023;13:1030693. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data used in this study were obtained from the Komodo Health Research Database, a proprietary claims and healthcare encounter dataset. Due to data use agreements and licensing restrictions, we are not permitted to share the raw data publicly. Researchers interested in accessing these data must obtain a license directly from Komodo Health (https://www.komodohealth.com/).