Mechanism-Based Therapy With Ampreloxetine for Neurogenic Orthostatic Hypotension in Multiple System Atrophy

Abstract

Background and Objectives

Degeneration of the central autonomic network with relative sparing of peripheral autonomic neurons underlies neurogenic orthostatic hypotension in patients with multiple system atrophy (MSA). Ampreloxetine, a novel, selective, norepinephrine (NE) reuptake inhibitor, allows once-daily dosing to precisely target residual peripheral autonomic neurons. Based on the hypothesis that patients with MSA would be most responsive and the substantial unmet need for symptomatic therapy in this population, an MSA subgroup analysis was prespecified.

Methods

We conducted a run-in 4-week, parallel-group, randomized controlled trial (SEQUOIA), followed by a pivotal enriched randomized withdrawal (RW) trial with 16-week open-label treatment and 6 weeks of 1:1 RW (REDWOOD). Inclusion criteria for the MSA subgroup included (1) probable or possible MSA, (2) 3-minute orthostatic blood pressure (BP) fall >20/10 mm Hg, and (3) dizziness or lightheadedness score >4 points. Outcome measures included self-reported symptom burden captured on the 10-item OH Questionnaire (OHQ). Differences were analyzed using logistic regression and mixed-model repeated measures analysis.

Results

Seventy-three patients with MSA entered the program (mean age 63 years old [range: 43–80], 52% male). Both SEQUOIA and REDWOOD did not meet their primary endpoints. In REDWOOD, 40 (61%) fulfilled enrichment criteria and were randomized. After 16-week open-label, OHQ symptom assessment (OHSA) composite domain scores improved 2.6 ± (SD = 2.1) points from pretreatment. The proportion of participants with treatment failure at week 6 of RW treatment period was 40% in the placebo arm and 15% in the ampreloxetine arm (p = 0.11). In secondary endpoints, at week 6 of RW, symptoms remained stable in the ampreloxetine group, but worsened on placebo (mean difference OHSA composite: −1.6 points ± 0.5; p = 0.0056, minimal clinically important worsening = 0.7–1.1 points). Standing for a short time favored ampreloxetine (−2.0 points ± 0.8; p = 0.015). Standing BP remained unchanged from open-label in the ampreloxetine group (systolic: 5.6 ± 4.1; diastolic: 3.7 ± 2.9 [SE] mm Hg) but fell after placebo withdrawal (systolic: −10.0 ± 4.5; diastolic: −6.0 ± 3.1 mm Hg). The catecholamine profile was consistent with NE transporter inhibition. There were no observed increases in supine BP.

Discussion

In a prespecified subgroup analysis of MSA participants in the REDWOOD trial, patients randomized to placebo worsened, whereas those who were randomized to treatment maintained their open-label level of function.

Trial Registration Information

REDWOOD trial, NCT03829657; first submitted to registry January 10, 2019; first participant enrolled February 22, 2019. SEQUOIA trial, NCT03750552; first submitted to registry November 20, 2018; first participant enrolled January 24, 2019. See ClinicalTrials.gov for full-protocol and statistical analysis plan.

Classification of Evidence

This study provides Class III evidence that in patients with MSA who had symptomatic benefit on orthostatic hypotension with ampreloxetine, there was no difference in the odds of treatment failures between those maintained on ampreloxetine and those withdrawn to placebo.

Introduction

Multiple system atrophy (MSA) is a rare, sporadic neurodegenerative synucleinopathy characterized by autonomic failure, parkinsonism, and cerebellar features in varying combination.¹ Prevalence estimates range from 3.1 to 12.4 per 100,000 in the US population.² The disease is rapidly progressive and usually leads to death 5–8 years after the onset of symptoms. Few patients survive for more than 10 years.³

Pathologically, MSA is defined by the deposition of insoluble aggregates of misfolded alpha-synuclein in the nervous system. The hallmark of the disease is the deposition of alpha-synuclein in glial cytoplasmic inclusions, in contrast to the Lewy bodies seen in neurons of autopsied cases of patients with Parkinson disease (PD).⁴ The gross pathology of MSA involves degeneration of the striatonigral system, cerebellum, pons, olives, and central autonomic network with relative sparing of the peripheral autonomic neurons.^1,5 This manifests in patients having normal venous plasma norepinephrine (NE) levels, compared with other Lewy body diseases (including PD and pure autonomic failure [PAF]),^5,6 where there is substantial degeneration of peripheral autonomic neurons.⁷

Degeneration of the central autonomic network in MSA underlies neurogenic orthostatic hypotension (nOH) in patients with MSA, which leads to physical deconditioning, functional decline, frequent falls, disability, dependency, and reduced quality of life. Furthermore, nOH is associated with an increased risk of cognitive impairment and early death from all-cause mortality.^8-11 Up to 80% of patients with MSA will develop nOH in their lifetime. Despite adherence to nonpharmacologic interventions and the available pharmacologic treatments, many patients remain symptomatic. A pressing need exists for additional effective pharmacotherapies.

Ampreloxetine is a novel oral, selective, long-acting, NE reuptake inhibitor that reaches a steady state of NE transporter inhibition and allows for once-daily dosing to precisely target residual peripheral autonomic neurons.¹² Phase 2 studies showed 10 mg/d oral ampreloxetine raised venous plasma NE levels in patients with nOH caused by all synucleinopathies.¹³ These data, taken together with the predominantly central neurodegeneration with relative sparing of peripheral autonomic neurons in MSA, suggest that ampreloxetine could provide a mechanism-based therapy for nOH in this patient population. This reflects a pathophysiology-aligned rationale-based therapeutic model.

Two back-to-back phase 3 trials of ampreloxetine for the treatment of nOH in patients with synucleinopathies (SEQUOIA, NCT03750552 and randomized withdrawal (REDWOOD), NCT03829657) were conducted. Randomization was stratified by disease type in both trials, and a prespecified MSA subgroup analysis was performed to determine the effect of ampreloxetine on symptoms, activities, and cardiovascular autonomic measures. Here, we report the results of that analysis. REDWOOD was considered the registrational trial, and the primary endpoint was the proportion of participants with treatment failure at week 6 of the randomized withdrawal (RW) in a 22-week study.

Methods

Study Design and Oversight

We conducted a run-in phase 3, 4-week, placebo-controlled, parallel-group, randomized controlled trial (RCT) (SEQUOIA) followed by a pivotal phase 3, placebo-controlled, RW trial (REDWOOD). Both SEQUOIA and REDWOOD were sequential multicenter, outpatient double-blind, placebo-controlled, randomized trials of a 10 mg/d single oral dose of ampreloxetine conducted at 76 autonomic or movement clinical sites in 19 countries (eTable 1). Although the studies ran back-to-back, they were independent of one another. Eligible patients were screened and enrolled in the 4-week run-in double-blind parallel-group, randomized, controlled trial (SEQUOIA) and rolled over into REDWOOD, which included a 16-week open-label steady-state enrichment period, with responders continuing into a double-blind, RW period (RW; 6 weeks). We permitted de novo enrollments into REDWOOD; however, most participants rolled over into REDWOOD from SEQUOIA.

The first patient in SEQUOIA began receiving ampreloxetine or placebo in January 2019, and the first patient entered REDWOOD in February 2019. The last patient began receiving ampreloxetine or placebo in June 2021, and the final patient entered REDWOOD in September 2021. The trials were designed by the sponsor (Theravance Biopharma) and an independent advisory board. Theravance Biopharma provided ampreloxetine and matching placebo tablets, monitored the trial sites, performed the laboratory tests, and analyzed the data.

Participants

The trials enrolled participants aged 30 years or older with PD, PAF, or MSA and symptomatic nOH. The prespecified subgroup analysis included all participants with an established diagnosis of MSA according to consensus criteria (Figure 1).¹⁴ Additional entry criteria included (1) diagnosis of nOH defined as a sustained fall in blood pressure (BP) of ≥20 or ≥10 mm Hg (systolic/diastolic) within 3 minutes of 60° upright tilt,¹⁵ (2) a Montreal Cognitive Assessment (MoCA) score >23 points,¹⁶ and (3) seated plasma NE levels ≥100 pg/mL. Inclusion required an item 1 dizziness or lightheadedness score ≥4 points on the validated Orthostatic Hypotension Questionnaire (OHQ, range 0 no symptoms to 10 worst possible symptoms).¹⁷ Patients with severe diabetic neuropathy were excluded (for full-entry criteria, see eMethods). After screening, other pressor agents were withdrawn. Fludrocortisone was allowed at a stable-dosing regimen.

Figure 1. CONSORT Flow Diagram.

Patient flow for the multiple system atrophy (MSA) subgroup enrolled in SEQUOIA, which was a 4-week randomized control trial, and randomized withdrawal (REDWOOD), which was a randomized withdrawal (RW) study with 16 weeks of open-label treatment followed by 6-week RW. Most participants rolled over into REDWOOD from SEQUOIA, and 2 participants entered REDWOOD as de novo. After SEQUOIA did not meet the primary endpoint, REDWOOD was early terminated by the sponsor, which is indicated in the discontinuation reason as “sponsor-driven discontinuation.” The prespecified subgroup analysis of the REDWOOD RW period included 40 MSA participants who were randomized into the withdrawal phase. *Both patients who enrolled de novo into REDWOOD were terminated early in the open-label phase by the sponsor because of closure of the trial after SEQUOIA did not reach the primary endpoint and therefore were not included in the analysis set. AE = adverse event; FAS = full-analysis set; MSA = multiple system atrophy; OL = open-label; RW = randomized withdrawal; QD = once daily.

Procedures

Figure 2 depicts the trial design and visits. Approved participants were first randomized 1:1 to a once-daily 10 mg ampreloxetine dose or matching placebo. The randomization sequence was generated using permuted blocks and stratified by disease type. A central interactive randomization and trial supply management system was used to assign participants to treatment arms and manage dispensing, ensuring concealment and maintenance of the blind for participants, investigators, and study staff. The Unified Multiple System Atrophy Rating Scale (UMSARS) was obtained.¹⁸ Participants were instructed to take a single oral 10 mg dose of ampreloxetine or matching placebo with 8 ounces of water at the same time each morning. No food or fluid restrictions were imposed. Education on nonpharmacologic nOH management strategies was provided. Those who successfully completed the 4-week SEQUOIA trial, with 80% trial medication compliance, were rolled over into REDWOOD. The 16-week open-label steady-state period of REDWOOD included an enrichment design with participants required to show symptomatic benefit compared with baseline at 2 continuation visits: (1) at week 4, a ≥2-point improvement in dizziness or lightheadedness (OHQ item 1),¹⁷ and (2) at week 16, a dizziness or lightheadedness score of ≤7, allowing room for worsening following placebo withdrawal. To minimize dropouts, 1 rescue dose of midodrine per week (10 mg max) was allowed during the 6-week RW segment. Owing to the COVID-19 pandemic, the protocol was amended to allow for telemedicine visits conducted by a nurse supervised by the treating investigator.

Figure 2. Study Design.

The prespecified subgroup analysis was performed in participants enrolled with a diagnosis of multiple system atrophy (MSA) and symptomatic neurogenic orthostatic hypotension (nOH), approved by an Enrollment Steering Committee, who met the cutoff for preserved peripheral (sympathetic) autonomic neurons (plasma norepinephrine (NE) ≥ 100 pg/mL). Approved participants were enrolled in a 4-week double-blind, placebo-controlled, randomized clinical trial of ampreloxetine (SEQUOIA). After completion, participants rolled over into randomized withdrawal (REDWOOD), which included a 16-week open-label period, with 2 enrichment points (week 4 and week 16). Participants who met open-label enrichment criteria underwent randomized withdrawal (RW) with 1:1 allocation to placebo or ampreloxetine for 6 weeks. The analysis of the efficacy outcome measures was performed at the end of each double-blind phase. SEQUOIA was designed to show benefit of ampreloxetine over placebo using the Orthostatic Hypotension Questionnaire (OHQ) item 1. REDWOOD was designed to show worsening after withdrawal to placebo in a composite treatment failure endpoint (1 point worsening of OHQ item 1 and a 1-point worsening of global patient impression of severity). ABPM = ambulatory blood pressure (BP) monitor; BP = blood pressure; MSA = multiple system atrophy; NE = venous plasma norepinephrine; OHQ item 1 = OHQ item 1 (dizziness or lightheadedness); OL = open-label; R = randomization; RCT = randomized controlled trial; RW = randomized withdrawal; Tx = treatment.

Outcomes

The primary outcome measure in SEQUOIA aimed to capture treatment benefit using OHQ item 1 (dizziness or lightheadedness)¹⁷ on ampreloxetine compared with placebo. The primary outcome measure in REDWOOD was designed to capture treatment failure (i.e., worsening) in patients withdrawn to placebo, compared with those maintained on ampreloxetine treatment. Treatment failure was defined as the occurrence of both: (1) a ≥1-point increase (worsening) in symptoms of dizziness or lightheadedness (OHQ item 1) and (2) a ≥1-point increase (worsening) in the Patient Global Impression of Severity (PGI-S) scale.¹⁹

Key secondary efficacy outcome measures included individual items and composite scores from the OHQ scale.¹⁷ Exploratory endpoints included BP and heart rate (HR) measurements obtained from an active stand test (10-minute supine and 3-minute standing). Pharmacokinetics and pharmacodynamics were measured at baseline and at the end of the 4-week RCT phase of SEQUOIA.

Clinical laboratory safety tests included vital signs, 12-lead ECG, comprehensive metabolic panel, a complete blood count, and urinalysis. Because a primary safety concern in patients with nOH is exacerbation of supine hypertension, the incidence of supine BP was obtained in 2 ways throughout the trial: (1) from clinic semisupine (30-degree angle) BP readings and (2) from 24-hour ambulatory BP monitors worn before treatment, and weeks 1 and 3 in SEQUOIA. They were not worn in REDWOOD. Supine hypertension was graded according to published criteria.²⁰ The Columbia-Suicide Severity Rating Scale was obtained at key visits.²¹ The incidence and severity of adverse events (AEs) were captured at each visit. The severity of AEs was subjectively graded by the physician as mild, moderate, or severe. An independent data safety and monitoring committee reviewed all safety data. All AEs were coded to the preferred terms of the Medical Dictionary for Regulatory Activities (version 24.1). Deaths were adjudicated by an independent committee, who were unblinded to treatment assignment.

Statistical Analysis

The SEQUOIA and REDWOOD trials were designed with a minimum enrollment target of 40% for patients with MSA. Although the MSA subgroup analysis was prespecified, the trials were only statistically powered based on the full population analysis (PD, MSA, and PAF) and not for individual diagnoses. As a result, the subgroup analysis in patients with MSA was not powered to demonstrate statistical significance, and no formal hypothesis testing was planned for this subgroup. All p values presented here are unadjusted for multiple testing and should be interpreted descriptively.

Disposition is presented for all MSA participants enrolled across both studies according to CONSORT reporting guidelines.²² Baseline characteristics were established at entry (before study treatment). Given that the REDWOOD trial was specifically designed in agreement with the FDA to evaluate the long-term durability of ampreloxetine, the efficacy analyses presented here focus on the RW period of the REDWOOD study which assess the effect of ampreloxetine withdrawal in patients with MSA. The analysis was based on the REDWOOD full-analysis set (FAS). The results from the preceding SEQUOIA trial, which assessed the initial induction of response (based on the SEQUOIA FAS), are summarized in the supplementary material. Pharmacodynamic results are reported based on the FAS from the SEQUOIA study. The FAS for both studies included all patients with MSA who had undergone randomization and received at least 1 dose of their assigned regimen (i.e., intention-to-treat) and had at least 1 postbaseline OHQ symptom assessment (OHSA) item 1 assessment. Safety analyses for both studies are reported based on the safety population (i.e. patients who received at least 1 dose of study treatment).

The primary endpoint in the SEQUOIA trial, OHSA item 1 was analyzed using a mixed-model repeated measures (MMRM) analysis. The MMRM model included factors for treatment, visit, treatment-by-visit interaction, and the baseline value as a covariable. The primary endpoint in the REDWOOD trial, treatment failure at week 6 of the RW period, was analyzed using a logistic regression model. The association between treatment and outcome is expressed as an odds ratio with its corresponding 95% CI. Participants in REDWOOD with missing primary endpoint data were imputed as treatment failures. Changes in OHQ endpoints, BP, and HR were analyzed using a similar MMRM analysis to the SEQUOIA trial. Changes in catecholamines are presented descriptively using geometric means, geometric mean standard errors, and the geometric mean ratio to prior visit. All statistical analyses assessed for the effect of heterogeneity by site with the results remaining consistent. All analyses were performed using SAS software version 9.4 and SAS/STAT version 15.1 (SAS Institute, Cary, NC).

Standard Protocol Approvals, Registrations, and Patient Consents

The trials had ethical approval by Advarra IRB (registration number 00000971) and by regional ethics committees (eTable 2), followed ethical principles of Good Clinical Practice set by the International Council Tripartite Guidelines, and written informed consent was obtained before participation.

Data Availability

Theravance will share individual deidentified patient data in line with regulations. The study protocols and statistical analysis plans (SAPs) are available on clinicaltrials.gov for SEQUOIA (protocol; SAP) and REDWOOD (protocol; SAP). Most versions are available on request.

Results

Population

Results of the full synucleinopathy population are available in clinicaltrials.gov (SEQUOIA and REDWOOD). Neither study met with the primary endpoint. Overall, 73 participants with a confirmed diagnosis of MSA were enrolled into SEQUOIA (71 participants) and REDWOOD (2 de novo participants; 64 rolled over from SEQUOIA) studies (Figure 1). The FAS included 67 patients with MSA for SEQUOIA and 40 patients with MSA for REDWOOD (Tables 1 and 2).

Table 1.

Demographics and Clinical Characteristics at Baseline (Pretreatment)

Measure	SEQUOIA		REDWOOD
	All (N = 67)	Placebo (N = 20)	Ampreloxetine (N = 20)	All (N = 40)
Male, (%)	35 (52)	13 (65)	8 (40.0)	21 (52.5)
White, (%)	63 (94)	19 (95)	17 (85.0)	36 (90.0)
Not Hispanic or Latino, (%)	61 (91)	17 (85)	17 (85.0)	34 (85.0)
Never smoker, (%)	47 (70)	14 (70)	14 (70.0)	28 (70.0)
Age (y)	64 (43–80)	63 (51–76)	64 (46–76)	63 (46–76)
NE (pg/mL)	268 (18)	271 (38)	220 (21)	244 (21)
BMI (kg/m2)	25 (4)	25 (5)	25 (4)	25 (5)
OHSA #1 dizziness score	6.6 (1.6)	6.7 (1.5)	6.5 (1.8)	6.6 (1.6)
OHSA composite score	5.3 (1.7)	5.4 (1.8)	5.2 (1.5)	5.3 (1.7)
OHDAS composite score	6.4 (2.2)	6.4 (2.6)	6.0 (2.0)	6.2 (2.3)
OHQ total score	5.8 (1.8)	5.8 (1.9)	5.6 (1.6)	5.7 (1.7)
Supine SBP 10 minutes (mm Hg)	142 (21)	137 (25)	144 (22)	141 (23)
Supine DBP 10 minutes (mm Hg)	86 (14)	82 (16)	87 (12)	85 (14)
Supine HR 10 minutes (beats/min)	70 (10)	69 (6)	71 (10)	70 (8)
Standing SBP 3 minutes (mm Hg)	99 (25)	90 (23)	103 (25)	97 (25)
Standing DBP 3 minutes (mm Hg)	67 (17)	60 (16)	69 (17)	65 (17)
Standing HR 3 minutes (beats/min)	83 (14)	82 (13)	85 (15)	84 (14)
Weekly falls because of faintinga, (%)	19 (28)	4 (20)	6 (30)	10 (25)
HADS anxiety total score	6.4 (3.7)	5.3 (2.5)	6.1 (3.7)	5.7 (3.1)
HADS depression total score	7.4 (3.9)	6.6 (3.3)	7.0 (3.4)	6.8 (3.3)

Abbreviations: BMI = body mass index; DBP = diastolic BP; HADS = Hospital Anxiety and Depression Scale; HR = heart rate; NE = norepinephrine; OHDAS = Orthostatic Hypotension Daily Activity Scale; OHQ = Orthostatic Hypotension Questionnaire; OHSA = Orthostatic Hypotension Symptom Assessment; SBP = systolic BP.

Note. NE presented as geometric mean (geometric standard error), age as mean (range), other continuous measures presented as mean (SD), and categorical presented as n (%).

^aNonmotor Symptoms Scale (score domain 1, question 2, frequency ≥ often [1/week]). Baseline plasma NE levels acquired seated before randomization into SEQUOIA.

Table 2.

Clinical Features of Multiple System Atrophy Population Set and Stage of Disease at Entry

Domain	SEQUOIA	REDWOOD
	All (n = 67)	Placebo (n = 20)	Ampreloxetine (n = 20)	All (n = 40)
MSA subtype, (%)
MSA-C	22 (33)	9 (45)	7 (35)	16 (40)
MSA-P	44 (66)	11 (55)	12 (60)	23 (58)
Unknown	1 (<5)	0 (<5)	1 (5)	1 (<5)
Time since MSA diagnosis (y)a	1.2 (0.3–2.0)	1.3 (0.7–2.3)	1.4 (0.3–2.6)	1.3 (0.4–2.5)
Duration of nOH (y)	1.7 (0.6–2.9)	1.2 (0.7–2.8)	2.3 (0.5–4.0)	1.6 (0.6–2.9)
Prior use of any nOH drugb, (%)	30 (45)	10 (50)	11 (55)	21 (53)
Prior use of pressor agentc, (%)	18 (27)	6 (30)	7 (35)	13 (33)
Concomitant use of fludrocortisone, (%)	14 (21)	4 (20)	6 (30)	10 (25)
Use of dopaminergic medicationd, (%)	45 (67)	12 (60)	14 (70)	26 (65)
UMSARS part I	21 (8)	21 (8)	20 (5)	20 (7)
UMSARS part II	23 (10)	22 (10)	21 (7)	22 (9)
UMSARS part IV, (%)
Completely independent	6 (9)	2 (10)	1 (5)	3 (8)
Not completely independent	28 (42)	7 (35)	12 (60)	19 (48)
More dependent	12 (18)	3 (15)	5 (25)	8 (20)
Very dependent	20 (30)	7 (35)	2 (10)	9 (23)
Totally dependent	1 (<5)	1 (5)	0 (<5)	1 (<5)
Walking aid needede	47 (70)	15 (75)	13 (65)	28 (70)
Imbalance/fallsf	27 (40)	6 (30)	8 (40)	14 (35)
Dysarthriag	41 (61)	13 (65)	10 (50)	23 (58)
Dysphagiah	15 (22)	5 (25)	3 (15)	8 (20)
Urinary dysfunctionj	34 (51)	10 (50)	12 (60)	22 (55)
Gait impairmentk	50 (75)	16 (80)	15 (75)	31 (78)

Abbreviations: MSA = multiple system atrophy; nOH = neurogenic orthostatic hypotension; UMSARS = Unified MSA Rating Scale.

Note. Continuous measures are presented as mean (SD), categorical as n (%), and time/duration as median (IQR).

^aTime since diagnosis of MSA.

^bIncludes vasoactive agents and volume expansion.

^cIncludes direct and indirect vasoactive agents midodrine, droxidopa, etilefrine, acarbose, pyridostigmine, atomoxetine, and pseudoephedrine.

^dIncludes levodopa, amantadine, dopamine agonists (pramipexole, ropinirole, rotigotine, and piribedil), MAO-B inhibitors (rasagiline, selegiline, and safinamide), opicapone, or entacapone.

^eUMSARS question 7 score 2 or above (assistance/walking aid needed occasionally or more).

^fUMSARS question 8 score 2 or above (falling < once a week, more than a week, or cannot walk).

^gUMSARS question 1 speech domain ≥ 2 moderate or worse.

^hUMSARS question 2 swallowing ≥ 2 moderate or worse.

^jUMSARS question 10 ≥ 2 urgency frequency moderate or above requiring drug treatment or catheterization.

^kUMSARS part II question 14 score 2 or above (moderate, severe, or cannot walk).

Baseline Pretreatment Characteristics

Eligible patients with MSA were 63 years old on average (range: 43–80), 94% White, 52% male, and had a median disease duration of 1.2 years (IQR = 0.3 to 2.0) and a median nOH duration of 1.7 years (IQR = 0.6–2.9 years) (Tables 1 and 2). Clinical characteristics were consistent with mid-to-late-stage disease, with manifest diagnostic signs/symptoms of MSA, including a high presence of urogenital dysfunction, dysphagia, dysarthria, imbalance/falls, gait ataxia, rapid deterioration with early need for walking aid, and poor response to levodopa (Table 2).^1,14 Eighteen (27%) reported prior use of a pressor agent, and 37 (55%) were nOH treatment naïve. Fourteen (21%) were treated with a stable dose of fludrocortisone.

Before receiving ampreloxetine, the mean baseline OHQ item 1 dizziness or lightheadedness score was 6.6 (SD = 1.6). Mean systolic BP in the supine position was in the mild supine hypertensive range (Table 1). All patients met criteria for nOH at entry.¹⁵ Systolic BP fell 43 mm Hg within 3 minutes of standing or tilt, with only a minimal increase in HR (13 beats/min).

Four-Week Randomized Control Trial (SEQUOIA)

A total of 71 patients with MSA were randomized, and 3 MSA participants dropped out. Over the 4-week double-blind, placebo-controlled, randomized trial with 1:1 allocation, OHQ item 1 scores were not significantly different in the placebo-treated vs ampreloxetine-treated arms (eTable 3).

RW Trial (REDWOOD)

Open-Label

As shown in the CONSORT diagram, 64 patients with MSA rolled over from SEQUOIA into REDWOOD. Two patients with MSA entered REDWOOD de novo. Of the 66 participants with MSA who entered the REDWOOD open-label phase, 40 (61%) completed the 16-week steady-state treatment, met enrichment criteria for improvement in symptoms, and went on to randomize (Figure 1). The main reasons for dropouts during the open-label were failure to meet the enrichment continuation criteria (n = 8, with a <2 point improvement in OHSA item 1), AEs (n = 7), withdrawal by patients for other reasons including motor deterioration (n = 6), physician decision (n = 1), and other (n = 2). After SEQUOIA did not meet the primary endpoint, REDWOOD was early-terminated by the sponsor that resulted in 2 sponsor-driven discontinuations (n = 2). There were no differences in the baseline characteristics of the patients who met the enrichment criteria vs not. At the end of the 16-week open-label, the OHQ item 1 dizziness or lightheadedness score and the nOH Symptom Composite (OHSA) decreased by 3.6 and 2.6 points, respectively, from the pretreatment baseline values (Figure 3B).

Figure 3. MSA Subgroup Patient-Reported Outcome Measures.

(A) Forest plot indicates odds of treatment failure (top) and least squares mean difference in Orthostatic Hypotension Questionnaire (OHQ) self-reported symptoms scores (bottom) at the end of the 6-week randomized withdrawal (RW). (B) Longitudinal analysis of mean OHQ symptom assessment composite scores, after censorship for early withdrawal, in the subgroup of randomized patients with multiple system atrophy (MSA) (n = 40). Note that benefit (lower scores) observed in open-label was maintained in the MSA group randomized to ampreloxetine (n = 20), but lost (i.e., increasing back to baseline) in the group withdrawn to placebo (n = 20). (C) Orthostatic Hypotension Daily Activities Scale (OHDAS) item 1 scores capturing severity of interference of symptoms of neurogenic orthostatic hypotension (nOH) on standing for a short time, which showed a similar pattern with improvement on open-label, that worsening by the end of the RW in the placebo arm. LS = least squares; OHDAS = OH daily activity scale; OHQ = orthostatic hypotension questionnaire; OHSA = OH Symptom assessment; RW = randomized withdrawal; SEM = standard error of the mean; Tx = treatment.

RW

The FAS for the prespecified subgroup analysis of MSA included 20 MSA participants randomized to placebo and 20 randomized to continue 10 mg/d ampreloxetine (MSA, per protocol population). Two patients did not complete the RW because of early termination of the study by the sponsor (i.e., sponsor-driven discontinuations), giving 19 completers in each arm (n = 38, total). The clinical characteristics at baseline were balanced in the treatment assignment arms (Table 2). Rescue midodrine was used infrequently in the RW period, with no use in the ampreloxetine arm and only 3 patients (15%) using rescue midodrine in the placebo arm.

Primary Outcome

REDWOOD failed to meet the primary endpoint in all patients with nOH (OR 0.60, 95% CI 0.27–1.29, p = 0.20). Prespecified MSA subgroup analysis showed the proportion of participants with treatment failure at week 6 of the RW was 40% in the placebo arm and 15% in the ampreloxetine arm. The study was not powered to detect differences in just MSA participants and as a result was not statistically tested, the observed odds of treatment failure on ampreloxetine relative to placebo was 72% lower (OR 0.28, 95% CI 0.05–1.22, p = 0.11). Analysis of the individual components of the treatment failure endpoint at week 6 revealed 12 placebo patients (60%) and 7 ampreloxetine patients (35%) had a ≥1-point worsening of dizziness or lightheadedness, while 7 placebo patients (35%) and 6 ampreloxetine patients (30%) had a ≥1-point worsening on the PGI-S scale. This suggests that the apparent reduction in the odds of treatment failure in the ampreloxetine group was primarily driven by the dizziness or lightheadedness question.

Symptoms and Daily Activities

Analysis of the OHQ individual symptom scale items showed worsening on placebo (Figure 3A). This difference was greatest in the least squared (LS) mean difference of the 6-item OHSA symptom burden composite assessment at the end of the RW. Similarly, functional scores were maintained in the continued treatment group. This was most noticeable in Orthostatic Hypotension Daily Activities Scale (OHDAS) subscale items that captured short-term daily living activities (Figure 3C). Standing for a short time (OHDAS item 1) and walking for a short time (OHDAS item 3) both worsened on placebo, with the biggest difference over time observed in patients with lower disability scores. Walking for a long time (OHDAS item 4) was the only individual item not to favor ampreloxetine.

Mechanism of Action

As summarized in Table 1, at entry and before randomization in SEQUOIA, pretreatment venous plasma NE levels in the MSA cohort were 268 pg/mL (range: 74–1,014 pg/mL). After 4 weeks of randomization, the catecholamine pharmacodynamic profiles on ampreloxetine showed increased venous plasma NE concentrations compared with both pretreatment baseline and compared with placebo (Figure 4A). Plasma DHPG levels were lowered after 4 weeks on ampreloxetine by 17% (eTable 3).

Figure 4. Norepinephrine and Blood Pressure.

(A) Venous plasma norepinephrine (NE) levels (seated) at baseline and at the end of the 4-week double-blind randomized controlled trial (RCT) period of SEQUOIA. (B) Longitudinal analysis of blood pressure (BP) at 3 minutes of standing at entry (pretreatment baseline), throughout the 16-week open-label period, and in the randomized withdrawal (RW) in the analysis population set of REDWOOD. (C) Absolute systolic BP after 10 minutes of supine rest and after 3 minutes of standing at week 6 of the RW. (D) Absolute diastolic blood pressure (DBP) after 10 minutes of supine rest and after 3 minutes of standing at week 6 of the RW. (E) Absolute HR after 10 minutes of supine rest and after 3 minutes of standing at week 6 of the RW. ^†p values vs placebo are based on least squared (LS) mean difference in change between the start and end of RW. Note that the minimal effect of ampreloxetine on supine BP with lower standing blood pressures and HR in the placebo group. DBP = diastolic BP; HR = heart rate; NE = norepinephrine; OL = open-label; RW = randomized withdrawal; SBP = systolic BP; SEM = standard error of the mean.

Figure 4B shows the longitudinal analysis of standing systolic BP throughout the REDWOOD trial. A pressor effect was observed on open-label treatment, with 3-minute standing BP increasing 6.5 mm Hg systolic and 5.3 mm Hg diastolic over the 16-week period compared with pretreatment measures. At the end of the RW, compared with open-label, standing 3-minute BP was lower in the group withdrawn to placebo (systolic: −10.0 ± 4.5, diastolic: −6.0 ± 3.1 mm Hg), while it remained stable in the group assigned to continue ampreloxetine (systolic: 5.6 ± 4.1; p = 0.016, diastolic: 3.7 ± 2.9; p = 0.029). As a result, at the final RW visit, BP after 3 minutes of standing was higher in the ampreloxetine-treated group (Figure 4, C and D).

Safety and Tolerability

The safety results are presented in Table 3. The most common AEs observed in the ampreloxetine group were urinary tract infections and headache. Because ambulatory monitors were not worn in REDWOOD, supine hypertension was determined by measurements made at 30°. Blood pressures obtained in clinic after 10-minute supine showed no changes on open-label treatment and in the RW phase in either arm (Figure 4C). During SEQUOIA and the RW of REDWOOD, there was only 1 AE for supine hypertension which was in the placebo arm during SEQUOIA. The number of falls was not different on/off treatment in the RW.

Table 3.

Safety Analysis

	SEQUOIA (4 wk RCT)		REDWOOD (6 wk RW)
	Placebo (n = 33), (%)	Ampreloxetine (n = 35), (%)	Placebo (n = 20), (%)	Ampreloxetine (n = 20), (%)
Participant with any treatment emergent
Adverse event	19 (58)	19 (54)	7 (35)	4 (20)
Moderate or severe AE	2 (6)	9 (26)	2 (10)	1 (5)
Severe AE	0 (<5)	2 (6)	1 (5)	1 (5)
AE related to study drug	5 (15)	5 (14)	1 (5)	1 (5)
Serious AE	2 (6)	4 (11)	1 (5)	2 (10)
Serious AE related to study drug	0 (<5)	0 (<5)	1 (5)	1 (5)
AE leading to permanent discontinuation	0 (<5)	1 (<5)	0 (<5)	0 (<5)
Death during study	0 (<5)	0 (<5)	0 (<5)	0 (<5)
Participants with at least 1 AE in >3%a	5 (15)	8 (23)
Headache	2 (6)	3 (9)
Urinary tract infection	2 (6)	3 (9)
Muscle spasms	1 (<5)	2 (6)

Abbreviations: AE = adverse event; MSA = multiple system atrophy; RCT = randomized controlled trial; RW = randomized withdrawal.

^aParticipants with at least 1 treatment-emergent AE, by the preferred term, which occurred at a frequency of >3% of patients with MSA.

Classification of Evidence

This study provides Class III evidence that in patients with MSA who had symptomatic benefit on orthostatic hypotension with ampreloxetine, there was no difference in the odds of treatment failures between those maintained on ampreloxetine and those withdrawn to placebo.

Discussion

In neither SEQUOIA nor REDWOOD was there a difference in the primary outcome between treatment and placebo groups or the subgroup with MSA. The results of the prespecified subgroup analysis of patients with MSA, shown in key secondary endpoints, show that participants randomized to withdraw to placebo worsened, whereas those who were randomized to treatment maintained their open-label level of function. The pharmacodynamic profile was consistent with the activity of a highly selective peripheral NE transporter inhibitor that increases bioavailability of NE by attenuating NE reuptake at the neurovascular junction,¹² with a consequent pressor effect²³ and possible positive inotropic and chronotropic effect.⁵

Although the primary outcome was not met in either the full population or the MSA subgroup, all outcomes measured by the OHQ (with 1 exception—walking for a long time) including the primary endpoint worsened after placebo withdrawal (Figure 3A). Specifically, the mean between-group differences in the OHQ total composite score, OHSA composite score, and OHDAS composite score all indicated a maintenance of an open-label treatment effect on ampreloxetine, which worsened in patients withdrawn to placebo. It is important that the magnitude of these observed differences met or exceeded the range of minimally important differences previously established through anchor-based analyses in the OHQ validation paper.²⁴

The difference observed between patients with MSA maintained on ampreloxetine compared with those withdrawn to placebo supports the hypothesis that NE reuptake inhibition is ideally suited to patients with central autonomic lesions, who have relatively intact peripheral autonomic neurons, and normal venous NE levels. Despite the REDWOOD trial being underpowered to show significance in the MSA subgroup analysis, the findings indicate a robust treatment response in this patient population lasting at least 22 weeks, achieved through selective NET inhibition, and support this mechanism-based approach to this specific nOH subpopulation.

A strength of this trial design is that it permits assessment of the durability of clinical efficacy—an important feature in MSA where most patients are likely to require therapy for several years. Of note, the accelerated approval for droxidopa was based on a primary endpoint of 1-week duration.²⁵ Furthermore, it is possible that the 16 weeks of open-label treatment in REDWOOD allowed participants sufficient time to establish treatment efficacy and detect a loss of response over 6 weeks of the withdrawal period. Detection of loss of response is an important consideration in a withdrawal design.

The results of this study highlight the challenges of conducting an nOH clinical trial in patients with MSA. Several items of the OHQ may also address cardinal nonorthostatic features of MSA related to motor impairment, which may limit their ability to capture the efficacy of an antihypotensive drug effect. Despite clarifying instructions on the OHQ, for example, “symptoms due only to your low blood pressure” and “symptoms that appear on standing,” discriminating OH symptoms from MSA symptoms unrelated to nOH is likely to be demanding for most patients. Walking for a long time was the only item not to show worsening on placebo withdrawal and is the item most likely to be confounded by the progressive extrapyramidal and cerebellar dysfunction in MSA. These challenges should be borne in mind in future nOH trials using the OHQ in central synucleinopathies, especially those in the rapidly progressive MSA.

The inclusion criteria in this study encompassed MSA patients with both possible and probable MSA, and the studied patients included patients across clinical disease stages.¹⁴ Reports suggest that the median time from disease onset to the requirement of ambulatory aids is approximately 3 years with wheelchair dependency at approximately 5 years.²⁶ Consistent with these reports, 70% of the MSA population enrolled in the ampreloxetine program required walking aids or assistance ambulating. Within this context, the results of this study suggest that by improving the ability to maintain an upright posture, the functional burden of the disease and quality of life could be improved in patients with MSA.

The worsening of symptoms after 6 weeks of placebo withdrawal and worsening in most functional measures of nOH was associated with a lower upright BP compared with the group maintained on ampreloxetine. Although the standing BP was different between groups, the effect of withdrawal was modest (Figure 4). A prior positive study reported similar pressor responses and poor concordance between the improvement in symptoms and BP changes.²⁷ The discordance is likely due to the temporal nature of the symptoms which were assessed by OHQ recall over 1 week vs the timing of the BP measurements at a single office visit. Blood pressure at any given time in a patient with nOH is influenced by multiple factors, including time of day, ambient temperature, volume status, and meals²⁸—all of which confound point-in-time hemodynamic measurements.²⁹ Nevertheless, ultimately the primary therapeutic aim when treating nOH is to improve daily function by reducing the burden of symptoms, and not on targeting a specific BP goal.^23,28 In this context, raising upright BP to levels just above the lower threshold for cerebral autoregulation may be sufficient to result in improvement in symptoms.

No worsening of supine hypertension was observed with ampreloxetine based on clinic measures (Figure 4, C and D). This finding is in contrast to studies with the direct α1-adrenoreceptor agonist midodrine and the NE precursor, droxidopa, in which supine hypertension was observed. The observation may be related to the mechanism of action of ampreloxetine, which increases the availability of physiologically released NE by inhibiting reuptake at the neurovascular junction, rather than direct stimulation of adrenoreceptors by circulating agonists. In support of this notion is the observation that the drug pyridostigmine, which enhances residual sympathetic outflow, also has a low prevalence of supine hypertension.³⁰ It is most likely that the pressor effect is due to vasoconstriction at the neurovascular junction; however, a contribution of NE-induced chronotropy and inotropy cannot be excluded (Figure 4E). This is a topic suitable for further study.

Measurements of venous plasma NE and metabolites lend support to the study rationale, effective target engagement, and the mechanism of action of ampreloxetine. First, the preservation of peripheral autonomic neurons is supported by the normal venous plasma NE levels in the patient cohort; second, venous plasma DHPG levels were lower compared with pretreatment baseline in the ampreloxetine group indicating reduced intraneuronal NE metabolism; and finally, the shift in the DHPG:NE ratio supports target inhibition of the NE transporter and reduced presynaptic reuptake at the neurovascular junction. Taken together, these neurochemical results provide the basis for the efficacy of ampreloxetine in this patient population.

There are several limitations to our study. The primary endpoint was not met in this subgroup analysis, and the positive findings are secondary endpoints. In the context of a failed overall trial and failed primary endpoint in the MSA subgroup, the interpretation of descriptive p values must be made with caution. Randomized withdrawal relied on treatment failure as a primary outcome, which equally weighted a 1-point worsening in both the 11-point OHQ item 1 and the 4-point PGI-S scale,^17,19 and may have low sensitivity to detect responsiveness in this rapidly progressive neurodegenerative disease.²⁹ The REDWOOD trial featured an enrichment design, which may limit the generalizability of the findings in the real-world setting. No difference in fall rates was observed during the randomized periods. Forthcoming pooled safety data analysis from future studies is needed to accurately define the rate of AEs in this fragile population. Imbalances between smaller subgroups entering the RW are inherent statistical factors in rare disease trials. A larger study sample may minimize the effect of these confounders and attenuate the variability. This RW trial currently provides Class III Level of Evidence. This is primarily based on the prespecified primary outcome of treatment failure, which was not different between groups. The challenge of subjectively measuring treatment efficacy in a rare disease remains a major hurdle for the field. The RW design remains the regulatory benchmark for showing durability in an nOH trial but raises the possibility that the blind may be incomplete because patients have been exposed to the drug and are aware of its effect.

In summary, these findings support the ongoing development of ampreloxetine as a targeted treatment for nOH and represent the first example of mechanism-based medicine in nOH and MSA. Based on the results of this prespecified subgroup analysis, a phase III trial of ampreloxetine focused on patients with MSA (CYPRESS NCT # 05696717) is in progress to confirm these findings.

Glossary

AE

adverse event

BP

blood pressure

FAS

full-analysis set

HR

heart rate

LS

least squared

MMRM

mixed-model repeated measures

MSA

multiple system atrophy

NE

norepinephrine

nOH

neurogenic orthostatic hypotension

OHDAS

Orthostatic Hypotension Daily Activities Scale

OHQ

OH Questionnaire

OHSA

OHQ symptom assessment

PAF

pure autonomic failure

PD

Parkinson disease

PGI-S

Patient Global Impression of Severity

RCT

randomized controlled trial

RW

randomized withdrawal

SAP

statistical analysis plan

Author Contributions

R. Freeman: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. H. Kaufmann: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. I. Biaggioni: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. V. Iodice: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. J. Jordan: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. R. Vickery: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. T. Geurin: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. M.J. Kmiecik: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. L. Norcliffe-Kaufmann: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data.

Study Funding

The studies and this manuscript were funded by Theravance Biopharma.

Disclosure

R. Freeman has received consultancy fees from Theravance Biopharma; H. Kaufmann has received consultancy fees and funding through academic grants for research support from Theravance Biopharma; I. Biaggioni has received consultancy fees from Theravance Biopharma; V. Iodice has received consultancy fees and funding through academic grants for research support from Theravance Biopharma; J. Jordan has received consultancy fees from Theravance Biopharma; R. Vickery has received personal compensation as an employee and a shareholder of Theravance Biopharma; T. Geurin has received personal compensation as an employee and a shareholder of Theravance Biopharma; M. J. Kmiecik received personal compensation as a contractor for Theravance Biopharma; L. Norcliffe-Kaufmann has received personal compensation as an employee and a shareholder of Theravance Biopharma. Go to Neurology.org/N for full disclosures.