Adjunctive retigabine in refractory focal epilepsy: Postmarketing experience at four tertiary epilepsy care centers in Germany

Epilepsy & Behavior 56 (2016) 54–58

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

Adjunctive retigabine in refractory focal epilepsy: Postmarketing experience at four tertiary epilepsy care centers in Germany R.D. Nass a,⁎, C. Kurth b, A. Kull c, W. Graf d, B. Kasper d, H.M. Hamer d, A. Strzelczyk c,e, C.E. Elger a, B.J. Steinhoff b, R. Surges a, F. Rosenow c,e a

Department of Epileptology, University Hospital Bonn, Bonn, Germany Epilepsiezentrum Kork, Kehl, Kork, Germany Epilepsy Center Hessen and Department of Neurology, Philipps-University, Marburg, Germany d Department of Neurology, University of Erlangen-Nuremberg, Germany e Epilepsy Center Frankfurt Rhine-Main, Goethe-University, Frankfurt am Main, Germany b c

a r t i c l e

i n f o

Article history: Received 2 November 2015 Revised 24 December 2015 Accepted 25 December 2015 Available online 30 January 2016 Keywords: Retigabine Ezogabine Trobalt Retigabin

a b s t r a c t Purpose: Retigabine (RTG, ezogabine) is the first potassium channel-opening anticonvulsant drug approved for adjunctive treatment of focal epilepsies. We report on the postmarketing clinical efficacy, adverse events, and retention rates of RTG in adult patients with refractory focal epilepsy. Methods: Clinical features before and during RTG treatment were retrospectively collected from patients treated at four German epilepsy centers in 2011 and 2012. Results: A total of 195 patients were included. Daily RTG doses ranged from 100 to 1500 mg. Retigabine reduced seizure frequency or severity for 24.6% and led to seizure-freedom in 2.1% of the patients but had no apparent effect in 43.1% of the patients. Seizure aggravation occurred in 14.9%. The one-, two-, and three-year retention rates amounted to 32.6%, 7.2%, and 5.7%, respectively. Adverse events were reported by 76% of the patients and were mostly CNS-related. Blue discolorations were noted in three long-term responders. Three possible SUDEP cases occurred during the observation period, equalling an incidence rate of about 20 per 1000 patient years. Conclusions: Our results are similar to other pivotal trials with respect to the long-term, open-label extensions and recent postmarketing studies. Despite the limitations of the retrospective design, our observational study suggests that RTG leads to good seizure control in a small number of patients with treatment-refractory seizures. However, because of the rather high percentage of patients who experienced significant adverse events, we consider RTG as a drug of reserve. © 2016 Elsevier Inc. All rights reserved.

1. Introduction Up to one-third of patients with epilepsy do not achieve seizure freedom with currently available anticonvulsant drugs. The first potassium channel modulator used in the treatment of epilepsy [1,2,3] was approved by the European Medicines Agency (EMA) in March 2011 under the name retigabine (RTG, Trobalt®) and by the United States Food and Drug Administration (FDA) in June 2011 under the name ezogabine (EZG, Potiga®) for use as adjunctive therapy for focal seizures in adult patients. After price negotiations between the German statutory health insurance fund and the manufacturer GlaxoSmithKline® failed due to a negative pricing prospect, RTG was withdrawn from the German market in June 2012 [4] but could still be imported from other European countries for those patients who had obtained good seizure control. The three pivotal, randomized, controlled clinical trials that led to the approval of RTG and one postmarketing, open-label, uncontrolled study included a ⁎ Corresponding author. E-mail address: [email protected] (R.D. Nass).

http://dx.doi.org/10.1016/j.yebeh.2015.12.034 1525-5050/© 2016 Elsevier Inc. All rights reserved.

maintenance period of no longer than 16 weeks [1,2,3,5]. Two longterm, open-label extension studies and a compassionate use program provided long-term results in some of the patients [6,7]. In addition to these controlled studies and programs, long-term postmarketing observations can provide further important insight concerning efficacy and tolerability under “real life conditions” [8,9]. Here, we retrospectively analyzed the clinical experience in patients treated with adjunctive RTG over a period of 4 years at four tertiary epilepsy centers in Germany. 2. Patients & methods We retrospectively collected data from in- and outpatients with drug-refractory seizures with focal epilepsy syndromes who began RTG treatment after its approval and introduction to the European market in May 2011. The data from each of the participating centers were collected separately and later anonymized and pooled. We excluded patients for whom no follow-up data were available. Patients who had previously participated in premarketing studies of RTG were excluded as well, since we chose to focus on use in a postmarketing environment.

R.D. Nass et al. / Epilepsy & Behavior 56 (2016) 54–58

After RTG was withdrawn from the German market in June 2012, patients who did not return for regular visits were called and informed of this decision. Advice was offered on how to either continue taking the drug or how to schedule visits for therapeutic reevaluation, if so desired by the patient. Data on concurrent anticonvulsants, seizure frequency, initiation and termination of treatment, and adverse events from electronic patient records and/or follow-up phone calls were extracted. In cases where precise reports of seizure frequencies or duration of treatment were lacking, estimates were used. For example, we used the first day of the month for treatment initiation or termination when no exact date was available. Numerical seizure frequencies were used when provided. For those cases in which no exact numeric frequencies were provided, estimates were used (n = 69). For example, if a patient record read “(…) had five seizures since the initiation of treatment 5 months ago (…)”, a seizure frequency of one per month was assumed. We divided the patients into six groups in an ordinal scale for further analyses. The six outcome groups consisted of (I) seizure-free; (II) response with 50–99% reduction in seizure frequency; (III) partial response with 25–50% reduction in seizure frequency or no change in seizure frequency but a cessation of severe and debilitating seizure types such as generalized tonic–clonic seizures and recurrent status epilepticus; (IV) indifferent response with a change of ±25% in seizure frequency; (V) aggravation with an increase of N 25% in seizure frequency or first onset of status epilepticus; and (VI) unclear response. This system enabled us to include patients in whom no exact numbers were documented, e.g. when a progress note read “(…) had no more grand mal seizures (…)”, a partial response was assumed or if it read “(…) used to have seizures every week, since starting RTG only had seizures every couple of months (…)”, a response was assumed. Patients for whom no follow-up data in seizure frequency existed were excluded from analyses of seizure outcomes. Adverse events were divided into several classes as depicted below. Descriptive and analytic statistics (normality tests, nonparametric, unpaired t-tests, contingency tables, Kaplan–Meier analysis, Cox regression analysis) were calculated using IBM® SPSS® Statistis Version 22 (IBM Corporation, USA) and Graphpad Prism® (Graphpad Software, La Jolla, California, USA). Paired samples were analyzed with Wilcoxon tests. The study was presented to the local ethics committee. Due to the entirely retrospective nature of the study, a full, formal audit was waived. 3. Results 3.1. Patient characteristics We identified 199 patients from four epilepsy centers. One patient, who was treated at two centers, was excluded. Three patients who were without sufficient follow-up data were excluded as well, leaving 195 remaining patients for analysis. The follow-up period lasted up to 4 years. Daily doses of RTG ranged from 100 to 1500 mg. The patients were started on RTG between May 2011 and June 2012. A summary of patient demographics is provided in Table 1.

55

Table 1 Summary of patient demographics (AED = antiepileptic drugs, RTG = retigabine).

Sex: • Male • Female Epilepsy syndrome: • Focal • Unclassified Invasive therapy: • Resection/vagus nerve/deep brain stimulation

n

%

115 80

59 41

151 44

77.4 22.6

77

39.5

Age, duration, comedication (range):

Mean ± SD

Median

• • • • • •

37.6 ± 13.3 years 14.2 ± 12.9 years 23 ± 12.8 years 8.2 ± 4.01 2.3 ± 0.8 701.8 ± 283.5

35 years 12 years 22 years 8 2 600 mg

Current age (12–72) Age at diagnosis (0–70) Duration of epilepsy (1–68) Number of previous AEDs (2–23) Number of AEDs combined with RTG (1–4) RTG dosage (100–1500 mg)

mostly discontinued RTG in the early titration phase or were early dropouts to follow-up, no quantifiable data on seizure frequency could be retrieved from the patient records. These patients were excluded from the abovementioned analyses on seizure frequencies. 3.3. Retention rate For our study, 193 patients were available for retention analysis; 22 patients were censored. Median retention lasted 229 days. The one-year retention rate amounted to 32.6%, the two-year retention rate was 7.2%, and the three-year retention rate was 5.73% (Fig. 1). According to a Cox regression analysis, the following factors were associated with longer retention: – Favorable seizure outcome (overall p = 0.011; partial response pb0.037, OR = 0.44; 95% CI = 0.2, 0.95; response p = 0.003, OR = 0.43, 95% CI = 0.25, 0.74, seizure freedom p = 0.091, OR = 0.28, 95% CI = 0.66, 1.22). The following factors were associated with shorter retention: – Occurrence of adverse events (p = 0.008 OR = 1.96, 95% CI = 1.19, 3.24). – Initiation of treatment 1–7 months (n = 39) prior to withdrawal from

3.2. Seizure outcome Eighty-four patients (43.1%) experienced no apparent effect on seizure frequency. In twenty-nine patients (14.9%), seizure frequency increased during the time period with adjunctive RTG treatment. Nineteen patients (9.7%) showed a partial response with a 25–50% reduction in seizures or freedom from severe seizure events such as status epilepticus or secondary generalized tonic–clonic seizures. Twenty-nine patients (14.9%) responded with a reduction in seizure frequency by more than 50%, and four patients (2.1%) achieved seizure freedom. The mean monthly seizure rate was significantly decreased from 45.6 ± 158.4 [median: 12; min: 0.16; max: 1800] before the initiation of therapy to 34.9 ± 97.3 [median: 10; min: 0; max: 1000] during treatment with RTG (p = 0.002). For those 30 patients (15.4%) who had

Fig. 1. Retention of RTG shown as a Kaplan–Meier plot (RTG = retigabine).

56

R.D. Nass et al. / Epilepsy & Behavior 56 (2016) 54–58

Table 2 Summary of adverse events. (UTI = urinary tract infection, SUDEP = sudden unexpected death in epilepsy patients).

Any AE Neuropsychiatric

Hepatic Urinary Vegetative Cardiorespiratory Weight Skin Death

Any adverse events Any neuropsychiatric adverse events Alertness, wakefulness, concentration or fatigue disorders Seizure aggravation, myoclonus Impaired balance, coordination, gait, tremor, dizziness Impaired speech Impaired cognition, memory, thought Depression, behavioral disorders, irritability, hallucinations Impaired visual acuity, double vision, retinal discoloration (n = 1) Sensory symptoms Impaired sleep Mildly elevated liver function test Urinary hesitation, retention, UTI, hematuria Pain, headache, nausea, diarrhea, sweating, sexual dysfunction Tachycardia, palpitation, conduction block, asystole, dyspnea Weight gain (n = 7), weight loss (n = 3) Rash, edema, urticaria, acral discoloration (n = 2) Possible SUDEP

the market versus initiation 7–14 months prior to withdrawal from the market (n = 152, p = 0.011, OR = 1.83, 95% CI = 1.15, 2.92).

No significant influence on retention time was found with regard to age of onset, current age, gender, number of previous AEDs, or the number of AEDs used in combination with RTG. At the end of May 2012, the withdrawal of RTG from the German market was announced and went into effect in June 2012. On June 1st 2011, 87 patients (44.6%) were still taking RTG. The one-year retention after June 2012 fell from 32.6% to 19.6%. The warning concerning discoloration was officially announced by the manufacturer on June 24, 2013. At this point, 15 patients were still on RTG. The one-year retention rate hereafter remained high at 63.5%. One hundred sixty cases (82.1%) with a reported, distinctive reason for discontinuing RTG were identified. The reasons to stop RTG were adverse events in 23.6%, lack of efficacy in 21%, both adverse events and lack of efficacy in 23.6%, aggravation of seizures in 5.6%, combination of adverse events and aggravated seizures in 5.6%, difficulty in obtaining RTG after the market withdrawal in 1.5%, and worrying about the discoloration warning in 1%. In the remaining 17.9%, no reason for the discontinuation was retrieved. 3.4. Adverse events From our cohort, 76.9% experienced adverse events. The most common adverse events were CNS-related problems such as fatigue, speech and language difficulties, and dizziness in 59% of patients. We observed urinary symptoms in 14.4%, mostly voiding problems and in some cases leading to recurrent urinary tract infections (UTIs) and significant retention. Discoloration was reported in three of twelve long-term responders during the second year of therapy. The majority of patients had discontinued RTG before the official letter warning of discoloration was released on June 24, 2013, and the topic received more scrutiny by treating physicians [10]. A list of adverse events in our cohort is shown in Table 2. 3.5. Blood and urine tests Serial laboratory values before and after the initiation of RTG were available for 59 patients. In 34 of these patients (57.6%), blood serum laboratory values showed no relevant changes. Mild elevations of liver enzymes occurred in 25 patients (25/195 in total, 22/58 of serial laboratory tested patients). This only led to a discontinuation of therapy in

N

%

149 115 53 38 37 29 21 16 15 3 1 25 28 16 5 10 5 3

76.4 59.0 27.2 19.5 19 14.9 10.8 8.2 7.7 1.5 0.5 12.8 14.4 8.2 2.6 5.1 3.1 1.5

three cases. Apart from the previously mentioned liver enzyme elevations, the following changes were noted: mild leukopenia in two cases, mild anemia in one, and mild elevation in lipase together with liver enzymes also in one case. None of these laboratory changes warranted further clinical interventions. Retigabine metabolites form crystals in urine that can cause false positive detections of bilirubin or urobilirubin [11]. Repetitive urine dip-stick samples before and after the initiation of treatment were obtained from 20 patients, 16 of which displayed elevated levels of bilirubin and/or urobilirubin (80%), while 4 patients (20%) did not develop this anomaly. None of the patients with bilirubinuria had abnormal liver function tests, and only two of them (12.5%) had urinary symptoms, roughly the same percentage of urinary problems observed in the entire cohort.

3.6. SUDEP and other causes of death Three patients in our cohort died from possible SUDEP at different time points after RTG was initiated (two men aged 37 and 38 years and one woman aged 60 years). All three were found dead at home, and no autopsies were performed. Two cases occurred within the first two weeks of treatment, while the third case occurred more than a year after the initiation of treatment. One of the deceased men reported emotional disturbances and a 70% seizure reduction during treatment; the other two patients died before their follow-up visits. A summary of the three SUDEP cases is given in Table 3.

Table 3 Clinical details of the possible SUDEP cases. Case 1 37-year-old male; 22 years of epilepsy with somatosensory auras and complex partial seizures; no cardiopulmonary risk factors; treated with RTG 300 mg/day, valproate 1900 mg/day, lamotrigine 200 mg/day, eslicarbazepine 1200 mg/day; VNS; no data on efficacy; found dead 9 days after initiation of treatment; no autopsy. Case 2 38-year-old male; 28 years of epilepsy with focal motor and secondary generalized tonic–clonic seizures; smoker; treated with RTG 900 mg/day, carbamazepine 700 mg/day, zonisamide 600 mg/day; had benefitted with an ~70% reduction in seizures; had reported aggression; found dead after 408 days of treatment; no autopsy. Case 3 60-year-old female; 26 years of epilepsy with aphasic and secondary generalized tonic–clonic seizures; hypertension; treated with RTG of unknown dose and valproate 2000 mg, levetiracetam 2000 mg/day, olmesartan 40 mg/day, hydrochlorothiazide 25 mg/day, trimipramine 50 mg/day, levothyroxine 75 μg/d; no data on efficacy; found dead 14 days after initiation of treatment; no autopsy.

R.D. Nass et al. / Epilepsy & Behavior 56 (2016) 54–58

3.7. Cardiac and ECG abnormalities Twelve-lead ECG data were available from a subset of 30 patients (15.4%). In 28 patients (93.3%), no new cardiac abnormalities were found. One patient ECG displayed a new left bundle branch block, and one ECG revealed a new bradyarrhythmia with self-resolving asystole. No QT prolongations were found. Three patients without ECG-data reported cardiopulmonary symptoms such as palpitations, tachycardia, and shortness of breath. One patient required cardiopulmonary resuscitation because an asystole developed after the intravenous administration of phenytoin within the context of a status epilepticus. The event was regarded as phenytoin-related rather than RTG-related. In summary, cardiopulmonary adverse events were observed in 2.58% of patients, and ECG abnormalities without symptoms were reported in 7.7% of available ECGs.

4. Discussion We report on the postmarketing use of RTG in patients with medically refractory focal epilepsies in four German tertiary care epilepsy centers. The results are similar to controlled clinical trials and recent postmarketing studies. Our cohort included patients who had epilepsy for many years and did not obtain seizure freedom after taking multiple drugs as well as – in about 40% of cases – after undergoing surgical and stimulation therapies. Whereas the patient demographics including age and duration of epilepsy and number of antiepileptic comedications as well as median seizure frequency before initiation of treatment were similar to those of randomized controlled clinical trials (RTCs) and open-label study extensions as well as to a recent postmarketing long term observation [1,2,3,5,7,9,12], the RTG-effects on seizure control were different. In the randomized controlled trials, dose-dependent responder rates (percentage of patients with a reduction of seizure frequency of more than 50%) amounted to about 30–50% [12] in the short term and N60% in the long term [7]. In contrast, only 14.9% of our patients achieved similar reductions in seizure frequency, which is in line with a recent postmarketing study [9]. Seizure-freedom was achieved by 7.1–11.5% of the patients during the open-label extension trials, whereas only 2.1% of our patients were seizure-free upon adjunctive RTG. Our long-term retention rate of 32.6% after 12 months was similar to that of other observational studies, whereas the two-year retention rate of 7.2% was lower than that of a long-term, open-label study (63.5%) and a recent postmarketing study analyzing people with cognitive impairment (20%) [7,8]. Adverse events, lack of efficacy, or aggravation of seizures (96%) was much more important for discontinuation than problems in obtaining the drug or concerns about discolorations. The withdrawal of RTG from the German market in June 2012 may still have had a modest effect on retention, since patients who started taking RTG 1–7 months or less before market withdrawal had a shorter retention time than patients who were started 7–14 months before withdrawal from the market. No new patients were started on RTG after June 2012. Retigabine was still obtainable via international pharmacies. The one-year retention after withdrawal from the market decreased from 32.6% to 19.6% but not to zero. This likely reflects that only patients with a clear benefit continued to take the drug. After the warning about discolorations was issued, 63.5% of the patients who were still taking RTG continued to do so for another year. Our interpretation is that patients who had clearly benefitted from treatment kept taking RTG even after the market withdrawal, while cases with equivocal treatment effect discontinued RTG. After all, the drug was still obtainable, if with a bit more effort and paper work for the treating physicians and the patients. Hence, we believe that it is fair to assume that the most important factors associated with retention were efficacy and adverse events, mostly affecting the CNS or urinary

57

system, though withdrawal from the market might have played a role as well. Types, frequency, and severity of adverse events basically replicated the findings of the RCTs and Phase II/III clinical development programs and open-label extensions as previously published [1,2,3,5,7,12]. The most common adverse events were related to central nervous dysfunctions such as fatigue, dizziness, or impaired speech. Urinary disturbance problems, a specific effect of RTG among anticonvulsant drugs, must be discussed in detail. While most cases of urinary retention in our cohort were mild, we also witnessed severe urinary retention that required temporary catheterization. We emphasize the manufacturer's recommendation that urinary safety must be monitored in patients taking RTG. It is assumed that urinary problems associated with RTG are reversible if the drug is withdrawn; however, the effects of long-term RTG exposure are not yet known [13]. For the portion of our patients with bilirubinuria on urine sample analysis, we think that these are false positive results because of the bilirubin-like RTG crystals found in the urine of RTG-treated patients. This “pseudobilirubinuria” did not seem to be related to adverse outcomes in urinary or hepatic health in our cohort. From our sample, blue-gray discolorations were observed in three patients. The risk of blue-gray discolorations accumulates with time, and the dosage and median time of their occurrence is thought to be 4.4 years, with N95% of discolorations occurring after two years of use (four months to seven years) [6,10,14]. The areas affected most frequently are fingernails, toenails, and lips, but discoloration of the palate and larger skin areas has been reported as well. Ocular discolorations can affect the conjunctiva and retina. Discolorations affect approximately one-third of long-term patients [14]. They were detected neither in the clinical nor in the preclinical animal studies. A higher rate of discolorations in our sample could have been expected if the retention times would have been longer. The manufacturer as well as the EMA and the FDA now recommend ophthalmologic exams which include a slit lamp examination, visual acuity testing, and dilated fundoscopy before initiating RTG treatment and then following up every 6 months during treatment. The drug should be discontinued if visual changes occur. In those cases with skin discoloration but without affection of the eye, we recommend close monitoring and offering alternative drugs. One patient in or cohort with retinal discoloration continued the drug for another 10 months, however ultimately discontinued the drug since the long-term consequences of RTG-induced retinal changes are unknown. One patient with skin discoloration was switched to another drug after seizures had reemerged. Another patient with toe discoloration still receives RTG to the best of our knowledge. A similar practice has been used in other cases reported in the literature — many patients chose to stay on RTG despite discolorations. The occurrence of three possible SUDEP cases in our sample raises some concerns. In adults with intractable epilepsy, up to 9.3 SUDEP cases per 1000 person years can be expected [15,16]. In our cohort, SUDEP incidence rate amounts to about 20 cases per 1000 patient years (given a total observation period of 53,664 days for all patients). To our knowledge, this is the first signal towards an association between RTG treatment and SUDEP and should thus be examined in a broader context of pre- and postmarketing studies. Neither the dose ranging study 205 nor the recent postmarketing studies by Lerche and colleagues and by Huber and colleagues reported any SUDEP cases [3, 5, 8]. French and coworkers noted one death due to diabetic ketoacidosis in the treatment arm of the RESTORE-1 trial but no cases of SUDEP [2]. Brodie and colleagues reported one case of SUDEP in the treatment group of the RESTORE-2 trial and one case in their placebo group [1]. In their postmarketing observation, Wehner and coworkers noted two deaths: one from SUDEP and one from an unknown cause [9]. In the long-term open-label extension studies of RTG published by Gil-Nagel and colleagues, four deaths were reported in 556 patients during a 32month period [7]. Two were described as possible SUDEP cases, one as a probable SUDEP case, while the fourth was caused by multiple myeloma. Two more deaths were reported after data cutoff: one due to cardiac arrest, one due to asphyxia. Neither was thought to be related to RTG

58

R.D. Nass et al. / Epilepsy & Behavior 56 (2016) 54–58

treatment [6, 7]. A calculation of SUDEP rate per patient year in the broader context has not been conducted. There is no obvious explanation for a potential increase in SUDEP risk with RTG, but disturbances of cardiac excitability and conduction could potentially play a role. The molecular mechanism of action of RTG consists in the opening of potassium channels Kv7.2 and Kv7.3, thereby hyperpolarizing the cell. These potassium channels are found in neurons and smooth muscle cells but not in cardiomyocytes, where Kv7.1 channels prevail [17,18]. Hence, it was postulated that RTG had no effect on cardiac electrophysiology. In fact, some authors hypothesized that RTG may protect against certain cardiac arrhythmias through indirect effects [19]. In any case, bearing in mind the potential in-vivo effects of RTG on cardiac electrophysiology, serial ECG studies were included in Phase II and Phase III trials and the NCT01227902 study [1,2,3,5]. No significant ECG abnormalities were found in the studies 205, RESORE-1 and RESTORE-2, but moderate QTcprolongations were found in 8% of patients in the NCT01227902 study. The manufacturer recommends monitoring ECG changes in RTG-treated patients with known disorders of cardiac structure or conduction such as long-QT syndrome, especially when electrolyte disturbances or combinations of the drug with other potential QT-prolonging agents may occur. For our study group, ECG data were available only in some patients. The ECG alterations were observed in 2 of our patients but did not include alterations of QT intervals. Uncontrolled seizures are the most important risk factor for SUDEP [16], and the SUDEP cases in our cohort may simply be due to inefficient seizure control instead of potential cardiac sideeffects of RTG. 5. Limitations The retrospective study design and the collection of data from four distinct epilepsy centers bear some inherent significant confounders and limitations. Differences between centers included the intervals between follow-ups, nomenclature of seizure classification, adverse events, emphasis on various additional clinical data, standards in ECG, and laboratory monitoring. This, together with the lack of a standard protocol when starting RTG, resulted in serial laboratory and ECG data being collected only for a subset of patients. We did not subdivide our groups according to daily doses, since many patients did not take the usual 200 mg tid, 300 mg tid, or 400 mg tid, but instead took other dosages. Furthermore, many patients stopped the medication during titration at various doses. This high interindividual variability in the daily drug doses is commonly found in daily clinical practice and obviously depends on individual tolerability, seizure control, etc., but a more systematic analysis of such data is very difficult. Furthermore, estimates of seizure frequency on a numeric scale are relatively unreliable, especially in a retrospective analysis [20]. A division in outcome classes and an analysis of retention, as demonstrated here, can in part compensate for this source of error. Hence, despite the mentioned limitations, we do believe that postmarketing observational studies such as ours do play a role in elucidating the understanding of the efficacy and safety of drugs in “real life” settings [21]. 6. Conclusion In our uncontrolled long-term observation study of RTG, both the efficacy and tolerability data obtained from our patients were similar to that of two other recent postmarketing observational studies, but the results were not as favorable as the results of the RTCs and long-term extension studies. Our data confirms that RTG can lead to good seizure control for a selected number of patients with difficult-to-treat seizures. Because of the relatively high proportion of significant adverse events, we consider RTG a drug of reserve. Acknowledgments We would like to thank Mrs. Vera Marquardt for her help with the follow-up patient visits.

Conflict of interest The authors of this observational analysis did not receive funding or other support by any of the manufacturers or distributors of retigabine/ ezogabine. CEE received honoraria for consultancy, expert testimony and lectures from UCB Pharma, Desitin and Pfizer. RS has received speaker fees from Cyberonics, EISAI, Novartis and UCB Pharma. HMH has served on the scientific advisory board of Cerbomed, Desitin, Eisai, GSK, Pfizer and UCB Pharma. He served on the speakers’ bureau of Cyberonics, Desitin, Eisai, GSK, Ingelheim Boehringer, Novartis and UCB Pharma. BJS has received honoraria for consultancy, expert testimony or lectures by Desitin, Eisai, UCB and Viropharm. AS has received support and honoraria from Bayer HealthCare, Boehringer Ingelheim, Desitin, Eisai, Pfizer and UCB Pharma.

References [1] Brodie MJ, Lerche H, Gil-Nagel A, Elger C, Hall S, Shin P, et al. Efficacy and safety of adjunctive ezogabine (retigabine) in refractory partial epilepsy. Neurology 2010; 75(20):1817–24. [2] French JA, Abou-Khalil BW, Leroy RF, Yacubian EM, Shin P, Hall S, et al. Randomized, double-blind, placebo-controlled trial of ezogabine (retigabine) in partial epilepsy. Neurology 2011;76(18):1555–63. [3] Porter RJ, Partiot A, Sachdeo R, Nohria V, Alves WM. Randomized, multicenter, doseranging trial of retigabine for partial–onset seizures. Neurology 2007;68(15): 1197–204. [4] GSK steigt mit Trobalt® aus. http://www.deutsche-apotheker-zeitung.de/politik/ news/2012/06/01/gsk-steigt-mit-trobaltR-aus/7381.html. [5] Lerche H, Daniluk J, Lotay N, DeRossett S, Edwards S, Brandt C. Efficacy and safety of ezogabine/retigabine as adjunctive therapy to specified single antiepileptic medications in an open-label study of adults with partial–onset seizures. Seizure 2015;30: 93–100. [6] CHMP E: Committee for Medicinal Products for Human Use (CHMP): Trobalt (Retigabine). Procedure No. EMEA/H/C/001245/II/0014. European Medicines Agency; 2013. [7] Gil-Nagel A, Brodie MJ, Leroy R, Cyr T, Hall S, Castiglia M, et al. Safety and efficacy of ezogabine (retigabine) in adults with refractory partial–onset seizures: interim results from two ongoing open-label studies. Epilepsy Res 2012;102(1–2):117–21. [8] Huber B, Bocchicchio M. A retrospective evaluation of retigabine in patients with cognitive impairment with highly drug-resistant epilepsy. Epilepsy Behav 2015; 44:234–7. [9] Wehner T, Chinnasami S, Novy J, Bell GS, Duncan JS, Sander JW. Long term retention of retigabine in a cohort of people with drug resistant epilepsy. Seizure 2014; 23(10):878–81. [10] Einschränkungen für die Anwendung von Trobalt® (Retigabin) — Behandlung kann zu Pigmentveränderungen von Augengeweben, einschließlich der Retina, und der Haut, Lippen und/oder der Nägel führen. http://www.akdae.de/Arzneimittelsicherheit/RHB/ Archiv/2013/201306241.pdf. [11] Kristian B, Wachtmeister K, Stefan F, Forsgren L. Retigabine as add-on treatment of refractory epilepsy—a cost-utility study in a Swedish setting. Acta Neurol Scand 2013;127(6):419–26. [12] Porter RJ, Burdette DE, Gil-Nagel A, Hall ST, White R, Shaikh S, et al. Retigabine as adjunctive therapy in adults with partial–onset seizures: integrated analysis of three pivotal controlled trials. Epilepsy Res 2012;101(1–2):103–12. [13] Brickel N, Gandhi P, VanLandingham K, Hammond J, DeRossett S. The urinary safety profile and secondary renal effects of retigabine (ezogabine): a first-in-class antiepileptic drug that targets KCNQ (K(v)7) potassium channels. Epilepsia 2012; 53(4):606–12. [14] Clark S, Antell A, Kaufman K. New antiepileptic medication linked to blue discoloration of the skin and eyes. Ther Adv Drug Saf 2015;6(1):15–9. [15] Surges R, Sander JW. Sudden unexpected death in epilepsy: mechanisms, prevalence, and prevention. Curr Opin Neurol 2012;25(2):201–7. [16] Surges R, Thijs RD, Tan HL, Sander JW. Sudden unexpected death in epilepsy: risk factors and potential pathomechanisms. Nat Rev Neurol 2009;5(9):492–504. [17] Tatulian L, Delmas P, Abogadie FC, Brown DA. Activation of expressed KCNQ potassium currents and native neuronal M-type potassium currents by the anticonvulsant drug retigabine. J Neurosci 2001;21(15):5535–45. [18] Wickenden AD, Yu W, Zou A, Jegla T, Wagoner PK. Retigabine, a novel anticonvulsant, enhances activation of KCNQ2/Q3 potassium channels. Mol Pharmacol 2000; 58(3):591–600. [19] Qi Y, Wang J, Bomben Valerie C, Li D-P, Chen S-R, Sun H, et al. Hyper-SUMOylation of the Kv7 potassium channel diminishes the M-current leading to seizures and sudden death. Neuron 2014;83(5):1159–71. [20] Hoppe C, Poepel A, Elger CE. Epilepsy: accuracy of patient seizure counts. Arch Neurol 2007;64(11):1595–9. [21] Brewer T, Colditz GA. Postmarketing surveillance and adverse drug reactions: current perspectives and future needs. JAMA 1999;281(9):824–9.