The efficacy and safety of adjunctive perampanel for the treatment of refractory focal-onset seizures in patients with epilepsy: A meta-analysis

2.1 Search strategy

We (Yiming Li and Ya Zeng) independently searched PubMed, ScienceDirect, and Central Register of Controlled Trials (CENTRAL) from their earliest dates up to February 1, 2021. The final search string was “perampanel” [Mesh] or “3-(2-cyanophenyl)-5-(2-pyridyl)-1-phenyl-1,2-dihydropyridin-2-one” AND “refractory partial-onset seizures” [Mesh]AND randomized controlled trial [ptyp]. No other filters were used. Studies were selected with the following entry criteria: Phase III, randomized, double-blind, placebo-controlled, parallel-group design with 6-week observational baseline and 19-week double-blind treatment phase (6-week titration period and 13-week maintenance period). Participants who meet the following criteria are included: age > 12, diagnosed with focal-onset seizures (partial-onset seizures) at least two ASMs failures in the previous 2 years, at least five focal seizures in the 6-week baseline phase and was taking stable doses of 1-4 approved concomitant ASMs. Exclusion criteria were as follows: not English language studies, RCT articles using the same experimental data, articles without recording 50% reduction in the frequency of seizures, and articles using indeterminate drug doses. No PROSPERO registration number for the moment. The specific screening process is shown in Figure 1.

2.2 Outcome measures

The primary outcome was the responder rate, defined as the percentage of patients who experienced a ≥50% reduction in seizure frequency in the maintenance period relative to baseline in the European Union. Responder rate is a single point in the entirety of response and not a continuous variable. Thus, the 50% responder rate provides less information from all possible responses; it may be less sensitive. Increasing the gauge of responder rate to 75% showed more significant improvements compared with placebo. So, we defined a ≥75% reduction in seizure frequency and seizure free as the secondary outcome.

Safety outcomes were as follows: 1. the proportions of patients who drop out of the treatment for any reason; 2. all TEAEs related to PER (ie, dizziness, somnolence, headache, fatigue, upper respiratory tract infection, nasopharyngitis, gait disturbance, or irritability).

2.3 Data and assessment of the risk of bias

To ensure the consistency of the data collection of each study, we conform to the conditions of the study of the following information into a structured Excel data table: research characteristics (such as sample size, titration time, and maintenance time), the patient's characteristics (such as age, gender, etiology, duration of disease, and complications), and treatment regimen, concomitant ASMs, and clinical outcomes (evaluation forms from the Cochrane Manual were used to assess the quality of selection, performance, detection, attrition, and reporting biases for each qualifying study [2011]).⁹

2.4 Statistical analysis

To develop the meta-analysis, we used Cochrane Collaboration's Review Manager software (RevMan 5.4) to derive pooled effect estimates, such as odd ratios (OR) and associated 95% confidence intervals (CI). Statistical significance was assessed at a nominal α level of .05. The heterogeneity index I² determined the choice of fixed-effect model or random-effect model.

3.1 Study characteristics

Two hundred and twenty-four records were initially identified by searching the database and trial register. Four randomized controlled trials (RCTs) were retrieved for detailed evaluation.^10-13 The essential characteristics of the studies are provided in Table 1.

3.2 Risk of bias assessment

Four studies were multicenter, randomized, double-blinded, placebo-controlled, parallel-group trials, which resulted in a low risk of bias (Table 2).

3.3 Efficacy

We compared four dosage types of PER (2, 4, 8, or 12 mg) with placebo; pooled data from the four RCTs showed that PER 4, 8, 12-mg groups all had a superior response compared with the placebo group, and the PER doses of 8 and 12 mg appeared to be more effective than 4-mg dose of PER (8 mg: ≥50% reduction, 25.6% vs 35.5%, P = .002; ≥75% reduction, 12.4% vs 16.8%, P = .07; seizure-free, 3.5% vs 3.6%, P = .99; 12 mg: ≥50% reduction, 25.6% vs 36.0%, P = .002; ≥75% reduction, 12.4% vs 19.1%, P = .01; seizure-free, 3.5% vs 3.7%, P = .86). We compared the PER doses of 8 mg with 12 mg in three efficacy outcomes (≥50% reduction, 35.5% vs 36.1%, P = .84; ≥75% reduction, 17.8% vs 19.1%, P = .64; seizure-free, 3.5% vs 3.7%, P = .85), the data showed that there is no significant difference between the doses of 8 and 12 mg. The details of 50% reduction in seizure can be found in Figure 2. To sum up, the efficacy of different doses of PER is in the following order: 8 = 12 mg > 4 mg, the minimum effective dose of PER may be 4 mg/d as no statistically significant difference in the 2 mg/d dose compared with placebo.

3.4 Treatment withdrawal and adverse events

In all trials, 152 (9.52%) and 27 (4.3%) patients withdrew from the study due to drug-related TEAEs in the PER-added and placebo groups, respectively (OR 2.50, 95% CI 1.64-3.82; I² = 0%; P < .0001). The incidences of patients' withdrawal from the study due to drug-related TEAEs were higher in PER supplementation with 2, 8, and 12 mg compared with the placebo group (2 mg, 6.7% vs 3.8%, P < .02; 8 mg, 8.4% vs 4.4%, P < .004; 12 mg, 17.0% vs 4.6%, P < .00001). PER supplementation with 4 mg showed no significant difference (3.7% vs 3.6%, P < .92) (Table 3).

In addition, 12-mg PER compared to 8 mg had a higher proportion of trial withdrawal (8.7% vs 17.0%; P < .00001) and TEAEs resulting in dose reduction/discontinuation (18.5% vs 32.0%; P < .00001) without a significant increase in efficacy. To sum up, the safety of different doses of PER is in the following order: 4 > 8 > 12 mg. Two milligrams was not included in this comparison because 2 mg had no efficacy; thus, we considered the safety of 2 mg to be meaningless.

The incidence of treatment-related TEAEs was higher in the PER-added group (59.6% vs 37.9%; P < .00001) (Table 3), which was also related to dosage. TEAEs were higher in PER supplementation with 4, 8, and 12 mg compared with the placebo group (4 mg, 45.4% vs 30.7%, P = .0001; 8 mg, 61.2% vs 34.8%, P < .00001; 12 mg, 75.2% vs 37.0%, P = .0002). There was no significant difference in PER supplementation with 2 mg (37.2% vs 31.9%, P < .29) (Table 3).

There was no statistical difference in severe TEAEs (5.1% vs 5.1%; P < .88). The most common TEAE was dizziness and there is a significant difference between the PER-added group and the placebo group (29.1% vs 8.1%; P < .00001). In addition, the incidence of major TEAEs in the PER-added group and the placebo group was as follows: somnolence (16.0% vs 7.2%; P < .001), headache (8.6% vs 10.0%; P = .94), fatigue (7.2% vs 4.4%; P = .03), upper respiratory tract infection (5.0% vs 3.6%; P = .28), nasopharyngitis (8.4% vs 8.0%; P = .86), gait disturbance (3.8% vs 2.2%; 0.06), irritability (7.1% vs 2.3%; 0.003), rash (2.4% vs 1.1%; P = .31), nausea (3.3% vs 2.8%; 0.72), and falls (11.2% vs 6.6%; P = .12) (Table 4).