Phase 3 Trial of Interleukin-1 Trap Rilonacept in Recurrent Pericarditis


A total of 141 patients were assessed for eligibility, and 86 patients were enrolled in the trial (Fig. S2). Enrollment began on January 9, 2019, and concluded on January 17, 2020. Reasons for the exclusion of patients who did not meet eligibility criteria are provided in Table S2. The mean age of the enrolled patients was 44.7 years, and 57% of the patients were female. The predominant underlying cause of pericarditis was idiopathic (in 85% of the patients), with 15% of the patients having post – cardiac-injury pericarditis. Approximately half the patients were taking glucocorticoids at the time of the qualifying pericarditis episode.

Of the 86 enrolled patients, 61 patients had completed the run-in period and had undergone randomization before enrollment was stopped because of the accrual of the prespecified number of adjudicated primary efficacy end-point events. A total of 3 patients did not undergo randomization because they did not meet the prespecified clinical-response criteria, and 7 did not complete the run-in period because of adverse events (in 4), a decision by the investigator (in 2), or another reason (in 1). The additional 15 patients who were still in the run-in period were allowed to complete the run-in period but did not undergo randomization; they transitioned directly to the long-term extension period.

Demographic and Clinical Characteristics of the Patients at Baseline.

The demographic and clinical characteristics of the patients at baseline were balanced between the two randomized trial groups (Table 1). The median duration of rilonacept treatment, including the run-in period, was 9 months (range, 3 to 14). The mean (± SD) adherence to the trial regimen (the number of actual administrations divided by the number of planned administrations) was 98.7 ± 4.6% throughout the entire trial (run-in and randomized-withdrawal periods).

Run-In Period

Mean Numerical Rating Scale Scores for Pain and C-Reactive Protein Levels over the 12-Week Run-In Period.

Numerical rating scale scores for pain and C-reactive protein (CRP) levels as assessed by a central laboratory were recorded during the run-in period, during which all the patients received rilonacept. The mean pain numerical rating scale score and mean CRP level at the baseline visit differ from those recorded for the qualifying pericarditis episode; to allow for the completion of screening procedures, the investigator was permitted to treat each patient with standard-of-care medications temporarily during the interval between presentation with the qualifying episode and the baseline visit or trial enrollment. A 3-day rolling mean was calculated on the basis of nonmissing values ​​over each successive 3-day interval. In accordance with the protocol, pain was assessed daily with the use of a numerical rating scale (with scores ranging from 0 to 10 and with higher scores indicating greater pain severity). CRP was measured at baseline, on day 4, and at weeks 1, 2, 4, 6, and 12. At week 12, a total of 81 patients had assessments of the CRP level, but 2 of these patients had discontinued treatment before week 12; therefore, only 79 patients were considered to still be participating in the run-in period. 𝙸 bars indicate the standard error.

In the 86 patients who participated in the run-in period, rapid (after the first dose of rilonacept) and sustained reductions in the mean pain score on the numerical rating scale and the mean CRP level showed resolution of the enrollment-qualifying acute pericarditis episode (Figure 1). The median time to pain response was 5 days (95% confidence interval [CI], 4 to 6), and the median time to normalization of the CRP level was 7 days (95% CI, 5 to 8). The median time to the prespecified treatment response was 5 days (95% CI, 4 to 7).

Manifestations of pericarditis (pericardial effusion, pericardial rub, or electrocardiographic changes), when they were present at baseline, resolved by the time of randomization, except in one patient (who was assigned to the rilonacept group); this patient had a pericardial friction rub, which resolved by the end of the randomized-withdrawal period. The median time to rilonacept monotherapy was 7.9 weeks (95% CI, 7.0 to 8.1); all the patients who had been taking glucocorticoids discontinued them and transitioned to receive rilonacept monotherapy during the run-in period.

Randomized-Withdrawal Period

Time to the First Adjudicated Pericarditis Recurrence.

Curves for the time to the first adjudicated pericarditis recurrence in the randomized-withdrawal period are shown. Circles indicate the time of data censoring for reasons other than a primary efficacy end-point event (eg, a visit at the end of the randomized-withdrawal period). Overall, 2 patients (7%) in the rilonacept group and 23 (74%) in the placebo group had pericarditis recurrence. The median time to recurrence could not be estimated in the rilonacept group and was 8.6 weeks (95% CI, 4.0 to 11.7) in the placebo group.

Trial End Points, Assessed in the Randomized-Withdrawal Period.

When the randomized-withdrawal period closed on May 29, 2020, a total of 25 primary efficacy end-point events had accrued (3 events occurred during the period of time required to complete the closure of the trial). During the randomized-withdrawal period, there were too few recurrence events in the rilonacept group to allow for the median time to the first adjudicated recurrence to be calculated; the median time to the first adjudicated recurrence in the placebo group was 8.6 weeks (95% CI, 4.0 to 11.7). Rilonacept led to a lower risk of pericarditis recurrence than placebo (hazard ratio, 0.04; 95% CI, 0.01 to 0.18; P <0.001 by the log-rank test) (Figure 2 and Table 2). During this period, in the intention-to-treat population, 2 of 30 patients (7%) in the rilonacept group had a pericarditis recurrence event, as compared with 23 of 31 patients (74%) in the placebo group.

The two recurrence events in the rilonacept group were associated with temporary interruptions of the trial-drug regimen, of one to three weekly doses; one interruption was due to poor adherence to the regimen, and the other was due to an adverse event, myalgia, which resolved. In the rilonacept group, 1 of the 2 patients who had pericarditis recurrence events received bailout rilonacept. In the placebo group, all 23 patients who had pericarditis recurrence received bailout rilonacept. No patient who received bailout rilonacept had pericarditis recurrence during the remainder of the randomized-withdrawal period.

The findings with regard to the primary efficacy end point were consistent regardless of baseline glucocorticoid use (Fig. S3 and the Supplementary Results section). The concordance between the investigator and clinical-event-committee assessments of recurrence events was 96.2% (Table S3).

All three major secondary efficacy end points (assessed at week 16 of the randomized-withdrawal period) showed a benefit of rilonacept monotherapy in providing a sustained clinical response and reducing the symptoms of pericarditis (Table 2). Results of sensitivity analyzes at week 8 and week 24 were consistent with those at week 16, and analyzes to address missing data were also performed (Tables S4 through S7).


For patients who did not discontinue the trial regimen and who transitioned to the open-label extension period, the adverse events reported here are those that occurred between the first dose of rilonacept in the run-in period and the last visit of the randomized-withdrawal period for patients who did not discontinue the trial regimen and who transitioned to the open-label extension period. For patients who discontinued rilonacept during the run-in period (10 patients) or discontinued rilonacept or placebo during the randomized-withdrawal period (1 patient) or at the end of the randomized-withdrawal period (1 patient), data on adverse events continued to be collected for 6 weeks after the last dose of the trial regimen.

Adverse Events.

Five serious adverse events occurred during the trial. One serious adverse event (stroke due to carotid-artery dissection) occurred during the run-in period, and four serious adverse events occurred during the randomized-withdrawal period; these included palpitations after alcohol ingestion (in the placebo group), squamous-cell carcinoma (in the rilonacept group), and pyrexia and postoperative ileus (in 1 patient each; both after rilonacept bailout in the placebo group). Overall, the investigators reported adverse events in 74 of the 86 enrolled patients (86%) (Table 3 and Table S8). Four patients had adverse events leading to the discontinuation of rilonacept therapy; these events, all of which occurred during the run-in period, included alopecia, extrinsic allergic alveolitis, erythema, and systemic allergic reaction (hypersensitivity). There were no deaths during the trial.

Injection-site reactions and upper respiratory tract infections were the most common adverse events. During the run-in and randomized-withdrawal periods, injection-site reactions (all of mild or moderate severity) occurred in 29 patients (34%), all of whom were rilonacept recipients. Upper respiratory tract infection was reported in 7 patients (23%) who received rilonacept before bailout and in no patients who received placebo before bailout. All the upper respiratory tract infections were mild or moderate in severity.

At week 24, the mean low-density lipoprotein cholesterol level, assessed in patients who were in a nonfasting state, was higher with rilonacept before bailout than with placebo before bailout (124.8 ± 33.4 mg per deciliter [3.25±0.85 mmol per liter] vs. 111.7 ± 24.4 mg per deciliter [2.90±0.65 mmol per liter]); the mean triglyceride level, assessed in patients who were in a nonfasting state, was also higher with rilonacept before bailout (198.0 ± 105.8 mg per deciliter [2.24±1.20 mmol per liter] vs. 96.7 ± 34.0 mg per deciliter [1.10±0.40 mmol per liter]). Details of the injection-site reactions, upper respiratory tract infections, and lipid variables are provided in Tables S9 through S12.

Leave a Reply

Your email address will not be published.