Health-related quality of life in patients with recurrent pericarditis: results from a phase 2 study of rilonacept

Background Impact of recurrent pericarditis (RP) on patient health-related quality of life (HRQoL) was evaluated through qualitative patient interviews and as an exploratory endpoint in a Phase 2 trial evaluating the efficacy and safety of rilonacept (IL-1α/IL-1β cytokine trap) to treat RP. Methods Qualitative interviews were conducted with ten adults with RP to understand symptoms and HRQoL impacts, and the 10-item Patient-Reported Outcomes Measurement Information System Global Health (PROMIS GH) v1.2 was evaluated to determine questionnaire coverage of patient experience. The Phase 2 trial enrolled participants with active symptomatic RP (A-RP, n = 16) and corticosteroid-dependent participants with no active recurrence at baseline (CSD-RP, n = 9). All participants received rilonacept weekly during a 6-week base treatment period (TP) plus an optional 18-week extension period (EP). Tapering of concomitant medications, including corticosteroids (CS), was permitted during EP. HRQoL was assessed using the PROMIS GH, and patient-reported pain and blood levels of c-reactive protein (CRP) were collected at Baseline and follow-up periods. A secondary, descriptive analysis of the Phase 2 trial efficacy results was completed using HRQoL measures to characterize both the impact of RP and the treatment effect of rilonacept. Results Information from qualitative interviews demonstrated that PROMIS GH concepts are relevant to adults with RP. From the Phase 2 trial, both participant groups showed impacted HRQoL at Baseline (mean PROMIS Global Physical Health [GPH] and Global Mental Health [GMH], were lower than population norm average). In A-RP, GPH/MPH improved by end of base TP and were sustained through EP (similar trends were observed for pain and CRP). Similarly, in CSD-RP, GPH/MPH improved by end of TP and further improved during EP, during CS tapering or discontinuation, without disease recurrence (low pain scores and CRP levels continued during the TP and EP). Conclusion This is the first study demonstrating impaired HRQoL in RP. Rilonacept treatment was associated with HRQoL improvements using PROMIS GH scores. Maintained/improved HRQoL during tapering/withdrawal of CS without recurrence suggests that rilonacept may provide an alternative to CS. Trial registration: ClinicalTrials.Gov; NCT03980522; 5 June 2019, retrospectively registered; https://clinicaltrials.gov/ct2/show/NCT03980522.


Introduction
Pericarditis, or inflammation of the pericardium, has a variety of etiologies but is most commonly referred to as "idiopathic" [1,2]. The primary symptom of pericarditis is debilitating chest pain. Pericarditis is considered recurrent if symptoms and inflammation recur at least 4 weeks after an initial acute episode [1]. Recurrent pericarditis (RP) affects approximately 15-30% of patients who have an acute episode of pericarditis, and up to 50% of patients who experience one recurrence will experience two or more [1]. Empiric "off-label" therapy with nonsteroidal anti-inflammatory drugs [NSAIDs] and colchicine is often used successfully to treat the first pericarditis episode or initial recurrence [3]. Treatment options for patients with multiple recurrences, however, are limited, and there is a high unmet medical need for patients who have inadequate response (i.e., continued recurrence or incomplete symptom resolution) to, who cannot tolerate, standard therapy [1] or who have persistent underlying disease. Given that the cytokines interleukin-1 alpha (IL-1α) and beta (IL-1β) are implicated in RP etiology [2,4], rilonacept, an IL-1 α and IL-1 β cytokine trap, was evaluated in clinical trials for the treatment of RP [5,6] and is now the first treatment approved by FDA for RP.
While the impact of RP on patients' health-related quality of life (HRQoL) has been reported in the literature [7][8][9][10][11][12] and is thought to be due to the primary symptom of the condition (e.g., chest pain) and the resulting uncertainty and anxiety about new recurrences, impact on HRQoL has not been explicitly evaluated in previous clinical research. In addition, corticosteroids (CS), despite well-known warnings and precautions in patients with cardio-metabolic comorbidities, are widely used to treat RP [13], putting patients at risk for additional adverse events, including recurrence and steroid dependence [14]; comorbidities associated with chronic CS use may also lead to adverse impacts on HRQoL [15,16].
Both qualitative and quantitative research approaches were used to explore the HRQoL impacts experienced by adults with RP and how those impacts may change in response to treatment. Qualitative interviews were conducted with ten adults with RP to document the patient experience of RP symptoms and HRQoL impacts (known as concept elicitation interviews). Results from these interviews were used to develop a conceptual model of RP. In addition, a Phase 2 clinical trial of rilonacept for the treatment of RP included an HRQoL patientreported outcome (PRO) questionnaire as an exploratory endpoint. Therefore, the objective of these two streams of research is to evaluate HRQoL in patients with RP: specifically, to confirm whether the concepts assessed with the HRQoL PRO questionnaire used in the Phase 2 clinical trial are relevant to patients with RP (based on patient reports during the qualitative interviews) and to evaluate the effect of rilonacept treatment on physical and mental aspects of HRQoL.

Methods
The sections below describe the methodology used for both the qualitative interviews with patients and the clinical trial study design relevant to the current research objective.

Methods for qualitative patient interviews to develop a patient-centric conceptual model
To understand the patient experience of RP, one-on-one telephone interviews were conducted with adults with a confirmed diagnosis of RP.

Participants for qualitative interviews
The qualitative interview study was approved by a centralized independent review board (IRB); following approval, potentially eligible participants were identified from clinical sites through review of medical records. Key inclusion criteria for the qualitative study included: age 18 years or older and a clinical diagnosis of RP (either idiopathic or due to post-pericardiotomy syndrome, adult onset Still's Disease, or Dressler's Syndrome), defined as the first episode of acute pericarditis (as defined by the 2015 European Society of Cardiology Guidelines for the Diagnosis and Management of Pericardial Diseases) [1] followed by at least one pericarditis recurrence after a symptom-free interval of at least 4-6 weeks. Key exclusion criteria for the study included: individual was currently enrolled in another clinical interventional study for RP; individual had a diagnosis of RP that was secondary to specific prohibited etiologies, including tuberculosis, neoplastic, purulent, or radiation, post-thoracic blunt trauma, myocarditis, or systemic autoimmune diseases (with the exception of adult onset Still's Disease).
Potentially eligible participants were presented with study information by a recruiting clinician (or his/her representative), and once participants provided a signed informed consent form, the clinical site completed a screening document to determine participants' eligibility. Participant interviews were scheduled once eligibility was confirmed.

Interview conduct
One-on-one, 60-min telephone interviews with ten adults aged 18-75 years with RP were conducted. Interviewers used a semi-structured interview guide to facilitate a conversational-style interview and included open-ended questions to understand the patient experience of RP and its treatments, specifically what signs, symptoms, and HRQoL impacts are experienced in relation to RP from the patient perspective.
Interview guide questions included:

Data handing and analysis
Interviews were audio-recorded after obtaining participant consent, transcribed, and anonymized. These transcripts were coded and qualitatively analyzed using the ATLAS.ti software program (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany). The goal of transcript coding was to organize and catalog participants' descriptions of the characteristics of RP, in order to develop a patient-centric conceptual model of RP signs, symptoms, and impact concepts. A conceptual model is a heuristic classification scheme that links a specified disease state or condition to its proximal and increasingly distal health outcomes [17], acts as a framework for understanding a disease and/or its treatment, specifies the potentially relevant outcomes for a program of research, and informs the selection of measurement concepts to foster the development of questionnaires, outcomes, and endpoints. Specifically, to characterize the specific applicability of the Patient Reported Outcome Measurement Information System Global Health (PROMIS GH v1.2) [17] questionnaire (described below) for capturing the HRQoL impacts experienced by patients with RP, conceptual mapping was conducted to compare the concepts within the conceptual model against PROMIS GH v1.2 individual items.

Methods for the Phase 2 study KPL-914-C001
The methodology of the Phase 2 clinical trial of rilonacept for the treatment of RP (clinicaltrial.gov: NCT03980522) is provided in detail by Klein and colleagues [5]. To summarize, this was a multicenter, open-label, single-activearm Phase 2 study that enrolled two specific patient populations of adults with RP: (1) patients with an active recurrence who were symptomatic and had signs of inflammation at baseline (A-RP) and (2) patients who were not having an active recurrence but were dependent on CS (CSD-RP). All participants received weekly subcutaneous (SC) injections of rilonacept for 6 weeks during the treatment period (TP) and were invited to continue weekly SC injections for up to 18 weeks in the extension period (EP).

Participants
Adults (18-75 years of age) with RP (idiopathic or postpericardiotomy syndrome etiology) were enrolled and stratified into one of two participant groups, A-RP and CSD-RP. Participants in the A-RP group either (1) had evidence of elevated c-reactive protein (CRP) at baseline or (2) did not have elevated CRP, potentially due to concomitant medications (such as CS) but had evidence of pericardial inflammation by cardiac magnetic resonance imaging. Participants in the CSD-RP group had CS-dependent disease (based on information from the investigator regarding prior recurrences when taking medication) and did not have active pericarditis symptomatology or elevated CRP at baseline.

Assessments
Blood levels of CRP were evaluated weekly in TP, and then monthly, to measure inflammation from baseline to the end of the EP for all participants. In addition, two PRO questionnaires were completed by all participants during the clinical trial. They included: 1. A single-item 11-point numeric rating scale (NRS) for average pericarditis pain intensity with a 24-h recall window (with 0 = no pain to 10 = pain as bad as it could be) [18][19][20] was completed weekly in TP and monthly in EP from baseline to the end of the EP. 2. The 10-item PROMIS GH v1.2 questionnaire was also completed by participants at up to five timepoints during the clinical trial to assess HRQoL [21]. The analyses presented here focused on the following three most critical timepoints: baseline (Day 0), end of TP (Week 6), and Final Visit (end of EP). Items 1-7 ask participants to think about their general health and are rated on a five-point response scale (with higher scores associated with better quality of life). Items 8-10 ask participants to report on the emotional problems, fatigue, and pain over the last seven days, with Items 8 and 9 rated on a five-point response scale (with higher scores associated with better quality of life) and Item 10 rated on a 0-10 NRS (with higher scores associated with more pain). Two domain scores are created from the 10-item scale, the global physical health (GPH) and the global mental health (GMH). The GPH is scored by averaging together the global03 (physical health), global06 (physical function), global07 (pain) and global08 (fatigue) items. The GMH is scored by averaging together the global02 (quality of life), global04 (mental health), global05 (satisfaction with discretionary social activities), and global10 (emotional problems) [21]. The published US-generalized normative scores for both of these domains are a mean score of 50 and a standard deviation (SD) of 10 [21].

Procedure and analyses
All participants received rilonacept SC injections weekly for 6 weeks until the end of TP and were invited to continue weekly SC injections (at the same dose) during an optional 18-week EP. For those on other concomitant medications for RP at baseline (participants in both the A-RP and CSD-RP groups), including CS, the option to taper was offered during the EP. Participants completed the PRO questionnaires (PROMIS and Pain NRS) at study visits, including telephone and site visits. Blood levels for CRP were also assessed during clinical site visits or via visiting nurse or local contract laboratory; CRP was analysed via a central laboratory.
The analyses presented in the results are descriptive, given the small sample size of the clinical trial and the single-active-arm design. Specifically, results are reported as means and SD, with ranges of values for each participant group. While participants were asked to complete the HRQoL questionnaire at multiple timepoints in the clinical trial, the analyses focus on the baseline, (Day 0), end of base TP (Week 6), and Final Visit at end of EP timepoints. Effect sizes (ESs; Cohen's d) were calculated with 95% confidence intervals (CI) to evaluate the magnitude of the change from baseline to end of EP for each patient group for the HRQoL scores. ESs ≥ 0.80 were considered large; ≥ 0.50 to < 0.80, medium; and ≥ 0.2 to < 0.5, small [22]. In addition, the descriptive analyses also include the weekly pericardial pain NRS scores and CRP blood levels that were collected at study visits.

Qualitative patient interviews
Qualitative interviews were conducted via telephone with ten adults diagnosed with RP to understand the patient experience of the condition, including the signs, symptoms, and HRQoL impacts. Participants were recruited from three clinical sites in the US. The mean age of the participants was 58.5 years (SD = 11.5), and six participants (60.0%) were female. Clinicians reported that participants exhibited the following RP types: idiopathic (n = 4, 40.0%), post-pericardiotomy syndrome (n = 4, 40.0%), adult-onset Still's Disease (n = 1, 10.0%), and Dressler's syndrome (n = 1, 10.0%). The majority of participants reported taking over-the-counter or prescription anti-inflammatory medications (n = 7, 70.0%), and half reported that they had previously taken CS (n = 5, 50.0%) for their RP.
In order to confirm that the assessment of HRQoL completed by participants in the Phase 2 clinical trial captured concepts relevant to adults with RP, concepts reported during the qualitative interviews were mapped to the ten items of the PROMIS GH v1.2 questionnaire. Table 2 shows the results of this exercise, with representative patient quotes from the qualitative interviews for each of the items of the PROMIS GH questionnaire. In particular, adults reported symptoms and HRQoL impacts during the interviews that are included in the PROMIS GH questionnaire, such as pain, social and emotional impacts, and physical functioning.

Phase 2 study rilonacept
Twenty-five participants were enrolled in a multicenter, open-label, single-active-arm Phase 2 clinical trial of rilonacept, with an average age of 42.8 ± 10.5 years (± indicates SD; range 26-62); most were female (n = 15, 60.0%) and white (n = 22, 88.0%). Participants had a mean number of prior recurrences of 2.6 (range 1-8), average duration of disease of 2.2 ± 1.9 years (range 0.2-7.9 years), and average number of pericarditis episodes per year of 3.9 ± 3.7 (range 0.54-15). Based on their baseline symptoms and signs of pericardial inflammation, there were two groups of participants: those experiencing an active recurrence who were symptomatic with evidence of inflammation (A-RP; n = 16), and those who were CS-dependent but not acutely symptomatic at baseline (CSD-RP; n = 9). See Table 3 for the demographics and health characteristics of these two participant groups. Scores from the PROMIS GH questionnaire items and domains were evaluated for each of the participant groups over time (baseline, end of TP, and end of EP). Figure 2 shows the baseline scores for the two domains of the PROMIS GH health questionnaire. For both the A-RP and CSD-RP groups, average scores for these domains are below the US normative average score of 50. Additionally, Table 4 shows a trend for improvement in some item and domain scores for both the A-RP and CSD-RP groups.
For participants in the A-RP group, increases in the average scores for items of the PROMIS GH questionnaire that assess general health, quality of life, and physical health indicate improvement over the study period, with large ESs. In addition, for the A-RP group, the average score of the PROMIS GH pain item shows the largest decrease over the study period, with a mean score of nearly 5 on the 0-10 NRS at baseline, and less than 1 at end of EP (ES = − 2.69; 95%CI = − 3.66 to − 1.72). At baseline, both the physical and mental domain scores (GPH/GMH) were lower than the normative average of 50, but by end of TP mean scores for the GPH were above the US norm (and remained above at end of EP, with a large change [ES = 1.48; 95%CI = 0.65 to 2.23), and mean scores for the GMH were at the US norm (and remained at the normative average at end of EP).
For participants in the CSD-RP group, there were modest increases (improvements) on the PROMIS GH items assessing mental health, social activities and relationships, social activities and roles, and emotional problems, with ESs indicative of medium changes (Cohen's d between 0.5 and 0.8). The average scores for the other items did not change. Similar to the A-RP group, average scores for the CSD-RP group were also below the US norm for both the GPH and GMH domains at baseline. For the GMH domain, average scores were at the normative average at the end of EP. Figure 3 shows the change in the GPH and GMH domain scores over the study period, along with the trend between these HRQoL scores and patient-reported pericardial pain and serum marker of inflammation (CRP). For the A-RP group, pain scores and CRP levels decreased over the study period (change in pain scores from Baseline to end of EP was 4. Please note that participants who completed the EP and were taking CS at Baseline (all participants in the CSD-RP [n = 8], and 83.3% [n = 5/6] of participants in the A-RP) were able to taper and/or discontinue using CS by the end of the EP (i.e., the end of the study) without recurrence or pericarditis symptomatology (e.g., patient-reported pericardial pain) or inflammation (e.g., elevated CRP level).

Discussion
While a substantial negative impact of RP on patients' HRQoL has traditionally been assumed, to our

Relevant disease process*
Swelling and irritation of the membrane surrounding the heart

Pericarditis symptoms
Chest pain     GMH domain (Cohen's d 0.50 to < 0.80), while not large (which was expected, given the absence of acute pericarditis recurrences) were not insubstantial, likely due to the tapering and discontinuation of corticosteroids. Future studies should investigator further the impact of CS use on HRQoL, and how discontinuation of CS for patients with RP on targeted therapy impacts physical and emotional HRQoL. Taken together, these results show that RP negatively impacts patients' quality of life physically and emotionally, and that improvements in quality of life may be associated with improvement in disease symptomatology and a decrease in pericardial inflammation, in particular, while patients receive targeted treatment. Furthermore, results from qualitative interviews, where adults with RP spoke about the unpredictable nature of the condition, supported that pericarditis recurrences impact patient physical and mental health. Specifically, using the qualitative data, a conceptual model of RP was developed, and HRQoL concepts included in the concept model (e.g., ability to carry out daily activities, impacts on mood, and limitation on social activities) were mapped against the ten items of the PROMIS GH v1.2 questionnaire (included in the Phase 2 clinical trial of rilonacept), which demonstrate that this questionnaire is assessing concepts that are relevant to adults with RP.
Limitations include the small sample sizes for both the qualitative interviews and the clinical trial, the singleactive-arm design of the clinical trial, and the relatively short duration (24 weeks) of the clinical trial compared to the overall duration of this chronic disease. In addition, Table 4 PROMIS GH item and domain scores over time (mean ± SD), by participant group CI = confidence interval; EP = extension period; ES = effect size; GMH = Global Mental Health; GPH = Global Physical Health; PROMIS GH = Patient Reported Outcome Measurement Information System Global Health; TP = treatment period a For Items 1-9, scores range from 1 to 5 with higher scores indicating improvement, and for Item 10, scores range from 0 to 10 with lower scores indicating improvement. Scoring for Item 10 is adjusted when calculating the GPH. To calculate the GMH and GPH domain scores, raw scores are converted to standardized T scores, with a normative mean of 50 and standard deviation of 10 b ES is calculated from Baseline while the inclusion criteria for the qualitative interview study were intended to be similar to those of the clinical trial, they were less restrictive (i.e., adults interviewed did not experience as many recurrences as the participants in the clinical trial). Nevertheless, these results provide preliminary support for the importance of including a multidimensional assessment of HRQoL for future clinical research of RP. It is also important to consider that some HRQoL impacts may be dependent on age and gender, therefore, given the age range of the participants who completed the qualitative interviews and the Phase 2 clinical trial, the resulting conceptual model should be considered representative of adult RP.
Strengths include leveraging qualitative results to support the importance of the item-and domain-level scores of the PROMIS GH v1.2 questionnaire to adults with RP. The representative patient quotes help contextualize how participants may be interpreting each item of the PROMIS GH questionnaire. In addition, the means for the PROMIS GH domain scores at baseline in the rilonacept clinical trial provide evidence of the impact RP has on patients' HRQoL, as they are lower compared to population norm scores. These findings are consistent with other clinical studies reporting lower PROMIS GH questionnaire scores and associated impacts in physical, mental, and social domains in cardiac and vascular populations [23][24][25]. The increase in both the GPH and GMH scores over the course of the study for both the A-RP and CSD-RP, in conjunction with improvements and/ or stable pericardial pain scores and CRP levels, shows that HRQoL scores may also be responsive to treatment as the patient's condition improves, particularly when on a treatment that addresses IL-1 driven pericardial inflammation.

Conclusions
Given the anxiety associated with the unpredictability of recurrences and the exercise restrictions that patients are expected to adhere to following a diagnosis of RP, it is important to evaluate both emotional and physical impacts of the condition. As more clinical trials move to incorporate patient-centric outcomes to evaluate treatments not only in terms of resolution of a physiological indicator of disease but also to ensure Fig. 3 PROMIS GH domain scores, pericardial pain, and c-reactive protein levels over time by participant group. This figure shows the change in the GPH and GMH domain scores over the study period for the A-RP group and the CSD-RP group, and the trend between these HRQoL scores and patient-reported pericardial pain and serum marker of inflammation (CRP). For the A-RP group, pain scores and CRP levels decrease over the study period, while HRQoL scores increase. For the CSD-RP group, pericardial pain and CRP (low at baseline, as expected because these participants entered the trial while not in active flare) remain low over the course of the study even while tapering and discontinuing CS, while HRQoL scores increase over time. BL = baseline; CRP = c-reactive protein; D = day; EoEP = end of extension period; EoTP = end of treatment period; NRS = numeric rating scale; QoL = quality of life; SE = standard error; W = week