Cardiovascular safety of low-dose fenfluramine in Dravet syndrome: a review of its benefit-risk profile in a new patient population

Abstract

Objective: Dravet syndrome (DS) is a rare, treatment-resistant epilepsy syndrome for which current treatment regimens are often ineffective. Fenfluramine is currently in development for treatment of DS, based on reports in the 1980s and 1990s of its anti-epileptic activity in pediatric patients with intractable epilepsy. However, fenfluramine was withdrawn from global markets in 1997 following reports of its association with pulmonary hypertension and heart valve disease in adult patients treated for obesity. This review was conducted to assess cardiac safety of fenfluramine when used at lower doses for treatment of DS.

Methods: Pubmed was searched for clinical studies of fenfluramine in obese adults who reported incidence of heart valve disease. These data were reviewed against published results from Belgian patients with DS who have been treated with low-dose fenfluramine for up to 28 years.

Results: Nine controlled studies of fenfluramine and related compounds (dexfenfluramine and/or phentermine) which assessed incidence and severity of cardiac valve disease in 3,268 treated patients and 2,017 control subjects have been reported. Mild or greater aortic valve regurgitation was found in 9.6% of treated patients compared with 3.9% of control subjects, and moderate or greater mitral valve regurgitation was found in 3.1% of treated patients and 2.5% of control subjects. Nineteen DS patients have been treated for up to 28 years with 10–20 mg/day fenfluramine, with no clinical signs or symptoms of cardiac valve disease or pulmonary hypertension. Slight and clinically unimportant changes in valve structure have been seen on echocardiography in five patients at some time during the observation period.

Conclusions: A different benefit-risk relationship appears to be emerging when fenfluramine is used at low doses for extended periods in young patients with DS. Continued cardiac assessments during ongoing Phase 3 clinical trials will provide additional safety information for this potential new and effective treatment.

Keywords:

• Epilepsy
• heart valve disease
• cardiac valvulopathy
• pulmonary hypertension

Introduction

Dravet syndrome (DS) is a rare, treatment-resistant epilepsy syndrome that has its onset in infancy^1,2 and is characterized by major motor seizures, developmental delay, cognitive impairment, and an elevated risk of mortality secondary to uncontrolled seizures and sudden unexplained death in epilepsy (SUDEP)^1–9. Current estimates of the incidence of DS range from one in 15,700 to one in 40,000, and more than 80% of patients have a mutation in the neuronal voltage-gated sodium channel, SCN1A^10,11.

Despite recent advances, treatment of DS remains challenging, and seizure control is often poor, partial, and transitory. A few conventional anti-epileptic drugs (AEDs) are useful in the treatment of DS (e.g. valproate, benzodiazepines, stiripentol, topiramate, bromides, and perhaps levetiracetam)¹², although no drugs are currently FDA-approved for treatment of DS in the US. Stiripentol has marketing authorization in Europe, Canada, and Japan¹². Other commonly prescribed AEDs (e.g. phenytoin, carbamazepine, and lamotrigine) are contraindicated, as they often worsen seizures in patients with DS¹². The majority of patients with DS are treated with a regimen of three or more AEDs, and yet many still experience frequent seizures and episodes of status epilepticus. Because of the disappointing efficacy of current AEDs, continuing research is being conducted to identify effective treatment for DS.

The appetite suppressant activity of fenfluramine (3-trifluoromethyl-N-ethylamphetamine, Figure 1) was identified in the mid-1960s¹³ and was approved in the US and globally for the treatment of adult obesity in 1973. However, serious cardiovascular adverse events (AEs) of cardiac valvulopathy and primary pulmonary hypertension (PPH) were recognized during its widespread use, typically at doses of 60–120 mg/day, and often in combination with phentermine^14,15. Subsequently, the drug was withdrawn from worldwide markets in 1997¹⁶. In the 1980s, investigators had conducted pilot studies of its use in epilepsy and other neurologic disorders in which reductions in seizure frequency were reported (for a review, see Schoonjans et al.¹⁷). Recently, the successful use of low-dose fenfluramine to control seizures in DS has been reported (see below)^18–20.

Figure 1. Molecular structure of ± fenfluramine.

Because of the noteworthy anti-seizure effect of fenfluramine that has been observed in patients with DS, it will be important to determine if the use of lower doses in this patient population confers a different benefit-risk profile as opposed to its use to treat obesity in adults. In the remainder of this paper, we will summarize past experience with the use of fenfluramine in the adult obese population as it pertains to cardiovascular safety, and then examine the preliminary cardiovascular findings from long-term use of low-dose fenfluramine in a Belgian cohort of patients with DS.

Methods

Data sources

Pubmed was searched for relevant studies of the association of the use of fenfluramine with cardiac valve disease or pulmonary hypertension. The following search terms were employed: fenfluramine, dexfenfluramine, cardiac valve, pulmonary hypertension, and adverse drug reaction. Review articles were also inspected for relevant studies. The data sources for the two cohorts of DS patients being treated with fenfluramine in Belgium are from all published reports, as well as the most recent update presentations at scientific meetings.

Selection of studies

Peer-reviewed reports of studies that described the prevalence of pulmonary hypertension or cardiac valve regurgitation were included in this review if sufficient detail was reported that described the patient population (e.g. percentage male or female), the primary drugs used (including combination therapy), the duration of therapy with fenfluramine or dexfenfluramine, the prevalence and severity of aortic and mitral valve regurgitation, or the prevalence of PPH. Case reports were excluded. In some cases, sequential reports from the same authors were presenting updates from the original patient cohort or from the original cohort plus additional patients. In these cases, only the most recent report was included.

Results

The most recent literature search returned 438 citations. Of these, four and 17 studies met the inclusion criteria for PPH and cardiac valve dysfunction, respectively.

Cardiovascular safety of fenfluramine in adult obesity

Fenfluramine was widely used as an appetite suppressant between 1973 and 1997¹⁶, often combined with phentermine. It was estimated that 1.2–4.7 million persons had been treated with fenfluramine, or its dextro isomer dexfenfluramine, in the US²¹. The initial report of cardiovascular AEs was made in 1981 by Douglas et al., who published the cases of two women who had presented with symptoms of pulmonary hypertension, including exertional breathlessness and exercise intolerance, while taking fenfluramine at doses of 120 or 160 mg/day²². In these two cases, the signs and symptoms of pulmonary hypertension resolved after stopping fenfluramine, but within several years a case of irreversible pulmonary hypertension associated with use of fenfluramine was reported²³. Two reports published in the New England Journal of Medicine brought the issue of cardiovascular AEs associated with the use of fenfluramine to the forefront. The first was a case-control study of 95 patients with PPH diagnosed between September 1, 1992, and September 30, 1994, who were compared with a cohort of 355 control patients matched for age and gender¹⁴. The use of appetite suppressant drugs, primarily dexfenfluramine or fenfluramine, was higher in the case patients (31.6%) compared with the control patients (7.3%; adjusted odds ratio for the risk of PPH = 6.3, 95% CI = 3.0–13.2). The use of these drugs for periods >3 months was associated with an odds ratio of 23.1 (95% CI = 6.9–77.7). The second report presented 24 cases of women with heart valve disease associated with the use of the fenfluramine-phentermine combination for weight loss¹⁵. Morphological changes were seen in both left- and right-sided valves with regurgitation in all cases. In this group of patients, the mean duration of treatment with the fenfluramine-phentermine combination was 11.0 ± 6.9 months (±SD), and the mean maximal doses of fenfluramine and phentermine were 56.5 ± 40.7 and 30.8 ± 8.5 mg/day, respectively. In addition, eight patients had echocardiographic or catheter evidence of pulmonary hypertension.

Shortly after the publication of these two reports, the US Food and Drug Administration (FDA) issued a public health advisory regarding the development of cardiac valvulopathy associated with exposure to fenfluramine or dexfenfluramine²¹. Within 3 months the FDA had received 144 reports, including the 24 index cases reported by Connolly et al.¹⁵, and, of these, 113 met the FDA case definition (documented aortic regurgitation of mild or greater severity and/or mitral regurgitation of moderate or greater severity after exposure to fenfluramine or dexfenfluramine). Nearly all cases were women (n = 111, 98%). The combination of fenfluramine and phentermine was used by 88 patients (79%), fenfluramine alone was used by two patients (2%), dexfenfluramine alone was used by 16 patients (14%), and all three drugs were used by six patients (5%). Based on five independent echocardiographic studies of patients who had been exposed to fenfluramine or dexfenfluramine alone or in combination with phentermine, the FDA estimated in 1997 that the prevalence of valvular disease was 33% (95% CI = 28–39%)²¹.

Primary pulmonary hypertension

After the publication of the initial report describing the association between the use of appetite suppressants and PPH¹⁴, others sought to determine the prevalence of this serious adverse effect. Rich et al.²⁴ reported the results of a survey of patients with PPH in North America. In their study of 579 patients with either PPH (n = 205) or secondary pulmonary hypertension (SPH, n = 367), they reported that the use of fenfluramine or dexfenfluramine for ≥6 months was associated with an odds ratio for PPH of 7.5 (95% CI = 1.7–32.4) compared with patients with SPH. Souza et al.²⁵ performed a retrospective chart review of 109 patients (95% female) with fenfluramine-associated PPH, and reported that the median duration of fenfluramine use was 6 months and the median time from start of fenfluramine to the onset of symptoms was 54 months (interquartile range = 24–116 months). Fenfluramine-associated PPH appears to have similar clinical, functional, hemodynamic, and genetic features, and overall survival as idiopathic or familial PPH²⁶. For example, Souza et al.²⁵ reported that the median duration of survival was 6.4 years in patients with fenfluramine-associated PPH, which was not different from that observed in patients with idiopathic or familial PPH, nor was it related to the duration of exposure to fenfluramine.

Cardiac valvulopathy

Many investigators performed a large number of both uncontrolled and controlled studies attempting to assess the prevalence of cardiac valve dysfunction in obese adult patients treated with fenfluramine and similar agents. Loke et al.²⁷ conducted a systematic review of randomized controlled clinical trials and selected observational studies. In 57 randomized controlled trials in which 2,907 patients had been treated with fenfluramine, dexfenfluramine, or phentermine for between 4–60 weeks (mean = 17 weeks), a single case of mitral regurgitation was reported as an AE and deemed to be due to a myocardial infarction, rather than related to fenfluramine treatment²⁷. No valve dysfunction was noted in 2,252 control subjects. It is important to note that routine echocardiography was not performed in any of the 57 randomized studies and, thus, asymptomatic valve dysfunction would not have been detected.

Table 1 presents a summary of uncontrolled echocardiographic studies of fenfluramine and related drugs originally summarized by Loke et al.²⁷, plus an additional study published after the Loke et al. review²⁸. In 7,022 treated patients in these studies, the prevalence of mild or greater aortic regurgitation was 16% (range = 6–29%), and the prevalence of moderate or greater mitral regurgitation was 5% (range = 1–11%). This summary suggests that about one in five adult patients treated with high-dose fenfluramine or related drugs for obesity were at risk of developing valvulopathy.

Table 1. Uncontrolled echocardiographic studies.

Reference (year)	n	Drugs	Daily dose (mg)	Mean duration (months)	Blinding	FDA AR (%)	FDA MR (%)
Burger et al.^45 (2001)	343	Fenfluramine-phentermine combination	Phe-FFA 15–30/20–60 (min–max)	12.3	No	6%	1%
Bowen et al.^21 (1997)	284	Fenfluramine, dexfenfluramine, and/or phentermine	FFA 40 Dex-FFA 30 Phe 30 (median)	14 (median)	No	28%	6%
Fisher and Ruden^46 (1998)	156	Fenfluramine-phentermine combination	NR	NR	No	19%	3%
Kancherla et al.^47 (1999)	200	Fenfluramine, dexfenfluramine, and/or phentermine	NR	10.2	No	12%	5%
Lepor et al.^48 (2000)	85	Fenfluramine-phentermine combination	FFA plus Phe 72 ± 27 (mean ± SD)	10.7	Yes	28%	11%
Teramae et al.^49 (2000)	191	Fenfluramine, dexfenfluramine, and/or phentermine	FFA 20–120 Dex-FFA 15–45 Phe 15–138 (min–max)	8.4	No	29%	6%
Wadden et al.^50 (1998)	20	Fenfluramine-phentermine combination	Phe-FFA 0–15/0–60 (min–max)	24.0	Yes	25%	10%
Dahl et al.^28 (2008)	5,743	Fenfluramine, dexfenfluramine (phentermine not specified)	NR	<6 mo, 36% 6–12 mo, 31% >12 mo, 33%	Mixed	15%	5%
Total	7,022					16%	5%

Adapted with permission from Loke YK, Derry S, Pritchard-Copley A. Appetite suppressants and valvular heart disease - a systematic review. BMC Clin Pharmacol 2002;2:6. https://bmcclinpharma.biomedcentral.com/articles/10.1186/1472-6904-2-6. Accessed April 7, 2017²⁷. The original article is published in Open Access: verbatim copying and redistribution of this article are permitted in all media for any non-commercial purpose, provided this notice is preserved, along with the article’s original URL.

FDA AR, FDA case definition for aortic regurgitation; FDA MR, FDA case definition for mitral regurgitation; NR, not reported.

An additional nine controlled studies were identified that sought to estimate the risk of valvulopathy in treated adult obese patients compared with baseline or with control groups of untreated subjects (Table 2). The prevalence of mild or greater aortic valve regurgitation at baseline or in untreated or placebo-treated subjects was 3.9% (range = 1.3–10.9%). In all studies, the prevalence of aortic valve dysfunction in treated adult obese patients was higher, 9.6% (range = 5.0–26.3%), but the difference was statistically significant only in four of the seven studies. Loke et al.²⁷ reported a relative risk of 2.82 (95% CI = 2.20–3.61) for aortic valve regurgitation in patients treated with fenfluramine and/or related compounds²⁸. Sachdev et al.²⁹ conducted a similar meta-analysis of controlled observational studies and found a similar odds ratio for aortic regurgitation of 2.2 (95% CI = 1.7–2.7%). Moderate or greater mitral regurgitation was present in 2.5% (range = 0.5–10.9%) of subjects at baseline or control subjects. In these nine studies, the prevalence of moderate or greater mitral regurgitation in treated patients was 3.1%. The results of these analyses of controlled studies indicated that fenfluramine-associated valvular disease was less common than the original observational reports suggested²⁹.

Table 2: Echocardiographic findings in controlled clinical studies of fenfluramine and related drugs in adult patients treated for obesity

Author (year)	Design	n	% Female	Drugs	Daily dose (mg)	Mean Duration(months)	Blinding	FDAAR (%)	FDAMR (%)	FDA ValveDisease
Gardin et al.^51 (2000)	PX	455	71.0	Dex-FFA	Dex-FFA 15-90	6.0	Yes	8.9**	4.9
		477	78.5	Phe-FFA	Phe-FFA 8-112/20-120	11.9		13.7***	5.1
		539	73.8	Control	–	–		4.1	3.2
					(range max daily dose)
Hensrud et al.^52 (1999)	PC	19	81.2	Phe-FFA	Phe-FFA 37.5/40	9.6	Yes	26.3	0
		11	36.4	Placebo	–	9.0		9.1	0
Jollis et al.^53 (2000)	PX	1,129	85	Phe-FFA	NR	11.1	Yes	8.9%***	2.6%
		669	74	Control	–	–		3.6%	1.5%
Khan et al.^54 (1998)	PX	233	87.1	Dex-FFA,Phe-Dex-FFA,or Phe-FFA	Dex-FFA 30,Phe-Dex-FFA 30/30,or Phe-FFA 30/60-120	20.5	Yes	NR, 21.9%^b	NR, 0.9%^b	22.7%***
		233	87.1	Control	–					1.3%
Shively et al.^55 (1999)	PC	223	81.6	Dex-FFA	NR	6.9	Yes	6.3%*	1.3%
		189	81.0	Control	–			1.6%	0.5%
Wee et al.^56 (1998)	L	46	74	FFA or Dex-FFAalone or with Phe	NR	5.3	Yes	13.0%	6.5%	15.2%^a
				Baseline				10.9%	10.9%	17.4%
Weissman et al.^57 (1998)	RDB	386	82	Dex-FFA	Dex-FFA 30	2.4	Yes	5.0%	1.7%	6.5%
		352	78	Dex-FFA (SR)	Dex-FFA (SR)	2.3		5.8%	1.8%	7.3%
		354	81	Placebo	30	2.3		3.6%	1.2%	4.5%
					–
Davidoff et al.^58 (2001)	FU of RC	276	100	FFA	FFA 60	≤3 (planned)	Yes	6.2%	5.1%
		254	100	Placebo	–			4.3%	4.7%
Ryan et al.^59 (1999)	L	86	20%	FFA or Dex-FFAwith Phe or Maz	FFA 60 or Dex-FFA 30with Phe 15 or Maz 1	16.7	Yes (not clear)	16.5%	5.1%
				Baseline				4.7%	4.7%
Total treated		3,268						9.6%	3.1%	–
Total control		2,017^c						3.9%	2.5%

* p < .05, **p < .01, ***p < .001 compared with control group.

Dex, dexfenfluramine; FDA AR, FDA case definition for aortic regurgitation; FDA MR, FDA case definition for mitral regurgitation; FFA, fenfluramine; FU, follow up; L, longitudinal; Maz, mazindol; NR, not reported; PC, prospective cohort; Phe, phentermine; PX, prospective cross-sectional; RC, retrospective cohort; RDB, randomized, double-blind; SR, sustained release.

^a 17.4% had FDA-valve disease prior to treatment. New or worsening valve disease was found in two (4.3%) of patients. Upon removing the eight patients with pre-existing valve disease from the analysis, new valvular abnormalities were seen in 2.6% of patients.

^b The original paper reported only total prevalence of valve disease. Loke et al.²⁷ reported the separate aortic and mitral regurgitation prevalence.

^c Only 1721 were evaluable from the data presented in the reports. In the Khan et al. study⁵⁴, only the percentage of patients in the control group with valvular abnormalities was reported (1.3%).

Other studies have demonstrated an increased risk of valvulopathy with longer duration or with higher doses of fenfluramine or dexfenfluramine and/or phentermine when used to treat adult obesity. Three studies have found an increased risk of valve dysfunction in these patients treated for >3–6 months compared with patients treated for ≤3 months^28,30,31. In addition to duration of exposure, evidence exists indicating that increased dose of fenfluramine confers a higher risk of valvulopathy. Li et al.³² examined the records of the patients in the original FDA report²¹, and found that, in patients taking the combination of fenfluramine and phentermine, the relative risk of severe valvulopathy was 9.2 (95% CI = 2.1–40.8) in patients taking ≥60 mg/day fenfluramine compared with patients taking <40 mg/day.

Assessing the magnitude of the risk of valvulopathy in obese adult subjects who were treated with fenfluramine or related drugs is confounded by the fact that many of the original reports made no effort to control for pre-existing valve disease, a finding that would not be unexpected in a population of obese adults with risk factors (i.e. the use of drugs with a serotonin [5-hydroxytryptamine, 5-HT]-based mechanism of action, history of rheumatic heart disease, etc.), and did not evaluate echocardiograms in a blinded manner^33,34. Mild or moderate valve insufficiency is usually asymptomatic and, when symptoms occur, typically dyspnea and/or reduced exercise capacity, they are similar to those observed in some otherwise healthy but obese subjects. Kimmel et al.³⁴ reviewed the medical records of 19 of the 24 subjects in the original report of fenfluramine-associated valve disease¹⁵, and found that four of the 19 (21%) subjects had pre-existing murmurs, four had symptoms suggestive of valve dysfunction prior to treatment, one had a history of rheumatic fever, and many had used drugs either known to be associated with cardiac valve disease or that had a 5-HT-based mechanism of action. The authors hypothesized that heart valve disease among fenfluramine users may be less common than previously estimated, that 5-HT excess may play a role in valve pathology, and that a patient’s response to anorexigens and other medications may serve as a marker for increased risk.

Cardiovascular safety of low-dose fenfluramine in Dravet syndrome

Two cohorts of Belgian patients with DS have been treated with fenfluramine under a Belgian Royal Decree that allowed the continued use of fenfluramine, because of the remarkable anti-convulsive activity observed with it, to treat patients with DS in the context of a clinical study following its market withdrawal. Patients in these cohorts have now been treated with fenfluramine for periods of up to 28 years^35,36. The original cohort began treatment before April 2010^18,19, and has been treated with fenfluramine for a mean duration of 17.2 years (range = 7–28 years)³⁶. Echocardiographic examinations have been conducted at least annually since 2010. The original cohort included 12 patients with a mean age of 8 years when fenfluramine was added to their current AED regimen, which included valproate for all patients, topiramate for six patients, and a benzodiazepine for six patients¹⁸. Patients were most often treated with 5 mg or 10 mg twice daily, for a mean dose of fenfluramine of 0.34 mg/kg/day. Two patients withdrew from the study; one due to lack of efficacy and one who remained seizure-free during and following a temporary treatment withdrawal due to a shortage of fenfluramine. In the initial report of this cohort, seven of the 10 remaining patients had been seizure-free for ≥1 year (mean duration = 6 years 7 months; range = 1–19 years)¹⁸. These 10 patients have been followed for an additional 5 years, and three were seizure-free for the entire 5-year period, and four others demonstrated seizure-free intervals of 2 or more years^19,36.

The most recent cardiovascular safety results for this cohort are presented in Table 3³⁶. None of the 10 patients who are still receiving fenfluramine has demonstrated any clinical signs or symptoms of heart valve dysfunction or pulmonary hypertension during their entire observation period^18,19,36. Slight changes in valve structure were detected at some echocardiographic examinations in five patients; however, at the most recent examination, no changes in valve structure were detected in eight of the 10 patients, including four who had echocardiograms read as having slightly thickened mitral and/or tricuspid leaflets at earlier examinations. Two patients have a mild stable thickening of one of the valves. Despite the slight thickened aspect of the leaflets, the valves do not show any restrictive or impaired function. These findings have been stable since 2010 and 2013, respectively. It should be noted that both of these patients are adults (31 and 26 years old) and in both the first abnormal echocardiographic findings were reported in only adulthood, 17 and 20 years after starting fenfluramine treatment, respectively. Unfortunately, no baseline echocardiography, to exclude pre-existing cardiac comorbidities, was performed in both patients. Valve thickening (although minor) is typically not seen in healthy adults (or children).

Table 3. Echocardiographic findings in the two Belgian cohorts of patients with Dravet syndrome treated with fenfluramine.

Patient number	Current age (years)	Years treated	FFA dose at last visit (mg/kg/day)	Number of echocardiographies performed	Previous echocardiographic findings (year)	Most recent echocardiographic findings of valvulopathy (year)	Clinical signs of valvulopathy or pulmonary hypertension
Original cohort
1	30.2	28	0.12	6	2012: slightly thickened AML without dysfunction; 2014: normal	2016: normal	No
2	41	25	0.26	9	2015: trace mitral regurgitation; no valvulopathy	2016: no valvulopathy; mild LV diastolic dysfunction (grade 1)	No
3	31	24	0.27	4	2010–2015: stable slight thickened aortic and tricuspid leaves without dysfunction	2016: stable slight thickened aortic + tricuspid leaves without dysfunction	No
4	26	22	0.33	7	2013–2015: stable slight thickened AML + tricuspid leaves without dysfunction	2016: stable slight thickened AML + tricuspid leaves without dysfunction	No
5	23.3	22	0.27	11	2014: slight thickened AML + tricuspid leaves without dysfunction	2016: Normal	No
6	28	15	0.20	7	Normal at all examinations	2016: no valvulopathy, trace mitral regurgitation	No
9	19	12	0.29	7	Normal at all examinations	2016: no valvulopathy, trace mitral regurgitation	No
10	21	10	0.24	7	2013: slight thickened AML without dysfunction; 2014–2015: normal	2016: no valvulopathy, trace mitral regurgitation	No
11	20	7	0.19	6	2010, 2013: slight thickened AML + tricuspid leaves without dysfunction; 2014: normal	2016: no valvulopathy, trace mitral regurgitation	No
12	8.8	7	0.42	9	Normal at all examinations	2016: normal	No
Prospective cohort
1	17.5	5.6	0.41	11	2010–2015: normal at all examinations	2016: subnormal systolic function due to mild hypokinesia of the interventricular septum^a	No
2	6.5	5.3	0.68	9	Normal at all examinations	2016: normal	No
3	7.4	1.5	0.61	6	Normal at all examinations	2016: normal	No
4	14.3	2.4	0.738	7	Normal at all examinations	2016: normal	No
5	15.4	1.9	0.26	7	Normal at all examinations	2016: normal	No
6	22.1	2.3	0.25	7	Normal at all examinations	2016: normal	No
7	21.9	1.6	0.27	5	Normal at all examinations	2016: normal	No
8	8.1	1.0	0.27	6	Normal at all examinations	2016: normal	No
9	30.6	0.8	0.24	5	Normal at all examinations	2016: normal	No

^a These findings were initially observed during an echocardiographic examination in April 2016 and were stable at the most recent echocardiographic examination in August 2016.

AML, anterior mitral leaflet; FFA, fenfluramine; LV, left ventricular.

This table is based on the most recent results presented during the American Epilepsy Society Annual Meeting^35,36.

The new cohort initiated treatment after April 2010 and has been followed prospectively with a structured protocol to assess safety and efficacy^20,35. The new cohort study began with a 3-month observation period to establish the baseline frequency of major motor seizures, as well as to provide a baseline echocardiographic examination²⁰. Echocardiographic examinations were conducted at 3-month intervals during the first year of treatment with fenfluramine, at 6-month intervals during the second year of treatment, and annually thereafter. At the time of the first report, nine patients had enrolled in the study (aged 1.2–29.8 years) and had been treated with fenfluramine, for a median of 1.5 years (range = 0.3–5.06). The initial mean dose of fenfluramine was 0.24 mg/kg/day. The frequency of major motor seizures had declined by a median of 75% from the baseline frequency of 15.0/month. Each patient had received five to 11 ongoing echocardiographic examinations, and no heart valve abnormalities have been found in any patient³⁵. At the start of Year 5 of treatment with fenfluramine, one patient demonstrated sub-normal systolic function (fractional shortening = 26% [normal = 25-43%] and ejection fraction = 53% [normal >55%]) without any clinical significance due to mild hypokinesia of the interventricular septum. This finding was stable at the follow-up echocardiographic examination performed 4 months later and was evaluated by two independent cardiologist observers. Both agreed that it was unlikely that it was related to the use of fenfluramine. No signs or symptoms suggestive of PPH have been observed in any patient.

Discussion

When fenfluramine was previously marketed for treatment of adult obesity, the emergence of serious cardiovascular AEs associated with its use and the use of related compounds (i.e. dexfenfluramine) resulted in a determination by the FDA that the benefits did not outweigh the risks of its continued use for weight loss. After its worldwide withdrawal in 1997¹⁶, a number of reports subsequently emerged in the literature that attempted to quantify the prevalence rate of cardiovascular side effects associated with the use of fenfluramine in this obese adult population. Initially these studies reported prevalence rates of over 25% in some cases (Table 1). Later controlled cross-sectional studies have suggested that the prevalence rate is more likely ∼12%. Regardless of these estimates, major challenges existed in interpreting the findings, based on a number of important limitations of the original studies, including lack of a baseline echocardiographic assessment to document any pre-existing heart valve disease and lack of control for pre-existing cardiovascular risk factors, such as a history of rheumatic fever or the use of medications known to be associated with valve disease. Despite these limitations, several studies found that dose and duration of therapy were correlated with development of valvulopathy in adult obese patients.

During the 1980s and 1990s, literature began to emerge that suggested that fenfluramine may be of benefit in treating refractory epilepsy in some patients. After the publication of a case series of 11 refractory epilepsy patients who were being successfully treated with low-dose fenfluramine³⁷, the Belgian investigators re-analyzed these patients and identified five with a positive genetic SCN1A mutation, which had recently been associated with DS³⁸. Based on the dramatic reduction in seizure frequency and long duration of seizure freedom, the investigators sought and received a Royal Decree in Belgium to allow them to continue to use fenfluramine in these patients with DS, even after worldwide withdrawal of the drug. Over the ensuing years, they have reported the outcomes of the two cohorts of patients that have been treated with low-dose fenfluramine, in some instances for over 25 years with continued success and without clinically relevant cardiovascular findings on echocardiographic examinations or clinical signs and symptoms of heart valve disease or pulmonary hypertension. Although the Belgian cohort started to use fenfluramine at a younger age (range = 1.2–29.8 years), 12 out of 19 reached adulthood (range current age = 6.5–41 years) in the meantime. Since even our adult patients tolerate long-term low-dose fenfluramine use well, the absence of clinically significant cardiovascular changes most likely cannot only be attributed to the younger age.

The pathophysiologic mechanism(s) responsible for fenfluramine-associated PPH or valvulopathy are not well understood³⁹. Yao et al.³⁹ have reported that fenfluramine results in dysregulation of multiple genes in cultured pulmonary artery smooth muscle cells that influence biological processes and pathways that are involved in normal function. Of interest was the presence of mutations in the gene encoding bone morphogenetic protein receptor (BMPR2), which was found in 22% of patients with fenfluramine-associated PPH, and was associated with a significantly lower duration of exposure to fenfluramine before developing PPH. These observations suggest that fenfluramine may act as a trigger or “second hit” for PPH in pre-disposed subjects²⁶. On the other hand, the pathophysiologic mechanism responsible for valvulopathy in subjects treated with fenfluramine is likely to involve 5-HT receptors. The mechanism of anti-appetite activity of fenfluramine is believed to involve central nervous system (CNS) neuronal 5-HT, as it is known that fenfluramine disrupts neuronal vesicular storage of 5-HT as well as inhibits the reuptake of 5-HT⁴⁰. The serotonin 2B (5-HT_2B) receptor has been implicated as the primary biochemical pathway leading to drug-induced valve disease³³. Its over-stimulation leads to proliferation and stimulation of valvular fibroblasts, with resultant collagen and glycosaminoglycan production, ultimately resulting in valve leaflet thickening. In addition, 5-HT and (+)- and (–)-norfenfluramine (the primary metabolite of fenfluramine) are potent agonists of the human 5-HT_2B receptor^41,42.

It is important to understand that the assessment of benefit-risk made in one patient population may not necessarily transfer to a different patient population. The same drug or treatment may have dramatically different benefit–risk ratios when used to treat different medical conditions; for example, fenfluramine used to treat adult obesity, compared with fenfluramine used to treat a devastating pediatric epilepsy syndrome like DS, which is associated with excess morbidity and mortality^1–3,7,43, and reduced quality-of-life for patients and their caregivers⁴⁴, and where a significant unmet medical need for therapies remains.

Of course cardiovascular safety remains an important issue for the use of fenfluramine in the treatment of DS, but, as of the most recent reports from the Belgian cohorts^35,36, no cardiovascular safety issues, including cardiac valve dysfunction or PPH, have emerged during up to 28 years of dosing. A different underlying risk profile for treatment with fenfluramine in children with DS as compared with obese, mostly female adults may be critical in this evaluation as well. The apparent threshold dosing duration for appearance of echocardiographic detectable cardiac valve dysfunction in obese adults is 3–6 months^28,30,31. Patients in the Belgian cohorts have been treated for a mean of 9.7 years (median = 5 years; range = 0.3–28 years). In addition, the doses used to treat DS have been lower than the doses that were used to treat adult obesity. The mean daily doses used in the Belgian cohorts were 5–20 mg/day, well below the threshold of 60 mg/day in obese adults that was associated with a 9-fold higher risk of severe valvulopathy compared with a dose of <40 mg/day³². It is, however, important to keep in mind that the exact dosage of fenfluramine when associated with valvulopathy was not always cited, and that different combinations (for example with phentermine or mazindol) had been used (see Tables 1 and 2). The combination of phentermine and fenfluramine in particular was very popular due to its synergistic effect in the treatment of obesity. It should be taken into account that phentermine might increase the effect (or side effects) of fenfluramine, making it difficult to compare the daily fenfluramine dose between different studies.

Conclusions

The observed cardiovascular safety in a group of patients with DS who have been receiving low-dose fenfluramine for a long period of time suggests that, unlike in the adult obese population, the benefits of the drug in DS treatment may outweigh the risks. Results from four ongoing global randomized, controlled Phase 3 trials will provide additional information that will help further define the benefit–risk profile of fenfluramine in DS.

Transparency

Declaration of funding

This study was funded by Zogenix, Inc.

Declaration of financial/other relationships

BC, AS, MP, MK, and LL have received honoraria for consulting and advisory board services from Brabant Pharma and Zogenix, Inc. FM and BP have nothing to disclose. AG is an employee of Zogenix, Inc. CMRO peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Previous presentations

The summary of the echocardiographic findings in patients with Dravet syndrome who have been treated with fenfluramine is based on data that were presented at the 2016 meeting of the American Epilepsy Society in Houston, TX.

Notice of correction

Please note that Table 2 has been corrected since the article was first published online (31 July 2017)

Acknowledgments

Professional medical writing and editing was provided by Edward Weselcouch of PharmaWrite (Princeton, NJ), which was funded by Zogenix, Inc. (Emeryville, CA).