Skip to main content
Free AccessScientific Investigations

Phrenic Nerve Stimulation for the Treatment of Central Sleep Apnea: A Pooled Cohort Analysis

Published Online: by:18


Study Objectives:

Early evidence with transvenous phrenic nerve stimulation (PNS) demonstrates improved disease severity and quality of life (QOL) in patients with central sleep apnea (CSA). The goal of this analysis is to evaluate the complete prospective experience with PNS in order to better characterize its efficacy and safety, including in patients with concomitant heart failure (HF).


Using pooled individual data from the pilot (n = 57) and pivotal (n = 151) studies of the remedē System in patients with predominant moderate to severe CSA, we evaluated 12-month safety and 6- and 12-month effectiveness based on polysomnography data, QOL, and cardiac function.


Among 208 combined patients (June 2010 to May 2015), a remedē device implant was successful in 197 patients (95%), 50/57 pilot study patients (88%) and 147/151 pivotal trial patients (97%). The pooled cohort included patients with CSA of various etiologies, and 141 (68%) had concomitant HF. PNS reduced apnea-hypopnea index (AHI) at 6 months by a median of −22.6 episodes/h (25th and 75th percentile; −38.6 and −8.4, respectively) (median 58% reduction from baseline, P < .001). Improvement in sleep variables was maintained through 12 months of follow-up. In patients with HF and ejection fraction ≤ 45%, PNS was associated with improvement in systolic function from 27.0% (23.3, 36.0) to 31.1% (24.0, 41.5) at 12 months (P = .003). In the entire cohort, improvement in QOL was concordant with amelioration of sleep measures.


Transvenous PNS significantly improves CSA severity, sleep quality, ventricular function, and QOL regardless of HF status. Improvements, which are independent of patient compliance, are sustained at 1 year and are associated with acceptable safety.


Fudim M, Spector AR, Costanzo M-R, Pokorney SD, Mentz RJ, Jagielski D, Augostini R, Abraham WT, Ponikowski PP, McKane SW, Piccini JP. Phrenic nerve stimulation for the treatment of central sleep apnea: a pooled cohort analysis. J Clin Sleep Med. 2019;15(12):1747–1755.


Current Knowledge/Study Rationale: Treatment options for central sleep apnea (CSA) are limited. Early evidence with transvenous phrenic nerve stimulation demonstrates improved disease severity and quality of life in patients with central CSA.

Study Impact: Transvenous phrenic nerve stimulation significantly improves CSA severity, sleep quality, ventricular function, and quality of life regardless of heart failure status. Improvements, which are independent of patient compliance, are sustained at 1-year and are associated with acceptable safety, thus supporting clinical use.


Central sleep apnea (CSA) is a breathing disorder characterized by diminished effort or cessation of breathing. CSA affects up to 40% of patients with heart failure (HF), irrespective of ejection fraction or cardiac symptoms1,2 and is a strong and independent predictor of morbidity and mortality.1,3,4 Further, CSA is often encountered in patients with atrial fibrillation, neurological disorders, and opioid use.5,6

CSA is caused by intermittent interruption of neural signaling from the respiratory control center, leading to a cessation of respiratory muscle activity and ventilation. The high prevalence of CSA in HF is closely linked to pathophysiological processes typical of HF, including exaggerated hypoxic and hypercarbic chemosensitivity, circulatory delay, altered cerebrovascular reactivity, and repeated apneic episodes, each associated with hypoxia and hypercarbia.7,8 CSA leads to heightened sympathetic tone, increased oxidative stress, systemic inflammation, and endothelial dysfunction.3,9,10

Treatment options for CSA are limited. Although positive airway pressure (PAP) is widely used to treat CSA, there are few data to support its efficacy in this sleep disorder. PAP is associated with inconsistent patient adherence and potential safety concerns, including increased risk of cardiovascular death in patients with heart failure and reduced ejection fraction with some forms of PAP therapy.8,11,12 The remedē System (Respicardia, Inc, Minnetonka, Minnesota, United States) is a transvenous unilateral phrenic nerve stimulator that restores physiological breathing patterns for the treatment of moderate to severe CSA.13,14 The pilot (n = 57, NCT01124370)15 and pivotal (n = 151, NCT01816776)16 trials that led to the US Food and Drug Administration (FDA) approval of the device in 2017 have been published. The comparable endpoints and follow-up requirements offer a unique opportunity to perform a pooled analysis of the two studies, conducted at different times and for the most part nonoverlapping sites, to provide clinicians with information on the largest number of patients with CSA treated with phrenic nerve stimulation (PNS) to date.


Pooled Cohorts

For the purposes of this analysis, we pooled patient level data from the two completed clinical studies evaluating the safety and efficacy of the remedē System for the treatment of CSA. The pilot study was a prospective, international, multicenter, nonrandomized feasibility, safety, and efficacy study at 13 hospital-based centers (5 in the USA, 4 in Germany, 3 in Poland, 1 in Italy). The study was conducted under a US FDA investigational device exemption ( NCT01124370).15 The pivotal study was a prospective, multicenter, randomized, open label controlled trial at 31 hospital-based centers (24 in the USA, 6 in Germany and 1 in Poland, with only 5 centers conducting both pilot and pivotal trials). The pivotal study was designed by members of the Steering Committee and the sponsor in consultation with the FDA ( NCT01816776).16 Details of the system, study design, and results have been previously published.1416 The respective protocols were approved by institutional or central ethics review boards, according to each center’s institutional procedures. The current analysis comprises all patients on active remedē System therapy including pivotal control individuals in whom PNS was initiated after the 6-month assessment of the primary endpoint (Figure 1). Effectiveness data through 12 months of active therapy were analyzed. The safety analysis was conducted in all patients with an implant attempt.

Figure 1: Pooled data set.

Twelve months of active treatment data were pooled from each individual trial. Only active treatment was considered and included in the timepoints in orange.

Patients were eligible for each original study if they met the following criteria during attended polysomnography (PSG) interpreted by a blinded core PSG laboratory: apnea-hypopnea index (AHI) ≥ 20 events/h of sleep, central apneas ≥ 50% of all apneas, and an obstructive apnea index of ≤ 20% of the total AHI. Key exclusion criteria were as follows: prohibitive factors for new device implantation, phrenic nerve palsy, Stage D HF, a percutaneous or surgical cardiac procedure ≤ 3 months, and advanced renal disease (creatinine level greater than 2.5 mg/dL, creatinine clearance ≤ 30 mL/min by the Cockcroft-Gault equation or on dialysis). Complete eligibility criteria for both studies are listed in Table S1 in the supplemental material.14 Patients were identified via in-home sleep apnea testing or review of prior clinically indicated PSGs. Prior to the baseline assessments, study patients had to have been medically stable for at least 30 days (no medication changes or hospitalizations) and be on guideline-recommended therapies appropriate for their clinical conditions. Potentially eligible patients underwent study-specific baseline overnight PSG within 60 and 40 days before system implantation in pilot and pivotal trials, respectively. Physiologic sleep data from both studies were interpreted by a core laboratory (Registered Sleepers, Leicester, North Carolina, United States) per The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications.17 Echocardiograms at baseline and follow-up were interpreted by core laboratories blinded to follow-up time and therapy duration in both studies (pilot, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania; pivotal, United Heart and Vascular Center, St. Paul, Minnesota).


All patients enrolled in the pilot and pivotal studies had an attempted remedē System device implantation, which consisted of a neurostimulator, stimulation lead, and a sensing lead (Figure 2). When possible, the device was placed on the right upper chest to accommodate an existing or potential future cardiovascular implantable electronic device. The stimulation lead was placed in the left pericardiophrenic or right brachiocephalic vein to stimulate the phrenic nerve, while the sensing lead was placed in a thoracic vein, such as the azygos vein, to sense respiration via thoracic impedance. In the pilot study and in the treatment arm of the pivotal trial, the device was activated at 1 month after implantation and therapy optimization was performed over the following 12 weeks using a proprietary algorithm that applied a stimulation pattern enabling full diaphragmatic contraction without disrupting sleep. The ranges of pulse stimulation used were 0.1–10.0 mA for 60–300 μs at 10–40 Hz.15 In the control group of the pivotal study, therapy was activated after the patients who had undergone treatment had their 6-month visit for the assessment of the primary endpoint. Follow-up visits were performed every 3 months until the end of the study period (up to 24 months) in pilot and until FDA approval in pivotal (median follow-up of 35 months). Follow-up assessments included attended PSG at 6 and 12 months.

Figure 2: Remedē System.

Copyright Respicardia, Inc, reproduced with permission.


Common endpoints at similar duration of follow-up in the pilot and pivotal studies15,16,18 allowed a safety and efficacy endpoint evaluation up to 12 months in the pooled data set. The primary efficacy outcome in pivotal was a comparison of the proportion of patients in the treatment versus control groups achieving a reduction in AHI ≥ 50% from baseline to 6 months. Although the primary endpoint in pilot was the change in AHI at 3 months, PSG tests were repeated in this study at 6 and 12 months. In order to perform this pooled analysis, we applied the criteria from pivotal to the following three groups: pivotal treatment group from baseline to 6 months, pivotal control group from 6 months to 12 months postbaseline while on active therapy (former control), and pilot patients from baseline to 6 months.

The primary safety endpoint in pivotal, defined as absence of serious adverse events related to device implantation, the system, or delivered therapy at 12 months, was also used as the primary safety endpoint for the pooled analysis. Prespecified secondary endpoints in the pivotal trial included: mean reduction in the central apnea index (CAI), AHI, and arousal index; mean increase in percentage of time spent in rapid eye movement (REM) sleep; the proportion of patients with moderate or marked improvement in the Patient Global Assessment health-related quality of life (QOL) instrument; mean reduction in the oxygen desaturation index of 4% or more (ODI4); and mean reduction in the Epworth Sleepiness Scale.19 These endpoints were also available in pilot and could be pooled. An independent data and safety monitoring board evaluated safety, and an independent clinical events committee reviewed and adjudicated all adverse events in each trial.

Statistical Analysis

The efficacy endpoint data were pooled based on months of active therapy and included the entire pilot group, the pivotal treatment group, and the pivotal control group after therapy was activated (Figure 1). The efficacy and safety endpoints were analyzed using the intention-to-treat population. Subjects missing an efficacy assessment were excluded from the analysis of that visit/endpoint. Missing data were not imputed. Similarity between pilot and pivotal baseline characteristics was assessed using analysis of variance for continuous variables and Fisher exact tests for categorical variables. Since this analysis is exploratory in nature, P values for statistical analysis of efficacy results were considered nominal, unadjusted for multiple tests. Binomial endpoints were tested using an exact test to assess if the proportion responding was greater than 0% (the Patient Global Assessment instrument tested the proportion of patients with marked or moderate improvement in QOL from baseline at each visit, and the New York Heart Association (NYHA) scale tested the proportion of patients with improvement in NYHA classification at each visit). Change from baseline to each visit for continuous endpoints was tested with paired t tests or the Wilcoxon signed-rank test based on the distributional characteristics of the endpoint. Tests were considered statistically significant if the nominal value of P < .05. Sensitivity analysis of the continuous endpoints was performed using repeated measures analysis of variance including terms for study, visit and a study by visit interaction; visit and change from baseline results were estimated from the model and P values for each model term were calculated to assess appropriateness of pooling and consistency of results.

The primary safety endpoint of freedom from serious adverse events related to implant procedure, device or delivered therapy through 12 months of follow-up was assessed in all study participants undergoing an implant attempt. Statistical analysis was performed using SAS (version 9.4, Cary, North Carolina, United States).


Overall, a total of 208 patients with an implant attempt were enrolled between June 2010 and May 2015 and were included in the pooled cohort analysis (n = 57 from pilot and n = 151 from pivotal) (Figure S1 in the supplemental material). At 6 months and 12 months, 170 (82%) and 156 (75%) patients, respectively, had data available for analysis.

Baseline Characteristics

Baseline characteristics were similar between both study cohorts except for smoking status, respiratory rate, left ventricular ejection fraction (LVEF) and mixed apnea index (Table 1 and Table S2 in the supplemental material). In the pooled cohort, comorbid HF (based on NYHA classification ≥ I per investigator assessment) was frequently present (68%) (73% of those with HF had EF ≤ 40% and 27% had EF > 40%). Patients had a high burden of comorbidities, including hypertension (74%) and coronary artery disease (58%). Both study cohorts had similar baseline PSG parameters such as AHI (pooled median [interquartile range (IQR)] 44.6 [32.7–59.4] events/h) and CAI (24.2 [15.0–39.9] events/h).

Table 1 Summary of baseline characteristics.

VariablePilot (n = 57)Pivotal (n = 151)PaPooled (n = 208)
Age (years)68 (60, 73)66 (59, 74).39966 (59, 73)
Male89 (51)89 (135)> .99989 (186)
Prior SDB therapy16 (9)37 (56).00431 (65)
Hypertension72 (41)75 (113).72474 (154)
Heart failure81 (46)64 (96).02068 (142)
NYHA classification (%).008
 No heart failure193632
Coronary artery disease61 (35)56 (85).53358 (120)
Atrial fibrillation28 (16)42 (64).07838 (80)
Left bundle branch block7 (4)9 (14).7849 (18)
Coronary artery bypass graft25 (14)24 (36)> .99924 (50)
Diabetes12 (7)30 (46).00725 (53)
Anemia5 (3)11 (16).2909 (19)
Cancer25 (14)15 (23).15418 (37)
Renal disease18 (10)25 (38).27423 (48)
Smoking status (%).003
COPD11 (6)11 (16)> .99911 (22)
Restrictive lung disease2 (1)3 (4)> .9992 (5)
Body mass index (kg/m2)29.0 (26.8, 32)30.2 (26.9, 34.6).03629.6 (26.9, 34.0)
Neck (cm)42 (40, 44)42 (40, 45).81742 (40, 45)
Heart rate (beats per minute)70.0 (64.0, 77.0)72.0 (64.0, 82.0).12671.5 (64.0, 80.0)
SBP (mmHg)120.0 (110.0, 135.0)124.0 (111.0, 135.0).840123.0 (110.5, 135.0)
Respiration rate (per minute)16.0 (14.0, 18.0)18.0 (16.0, 19.0)< .00116.0 (16.0, 18.0)
Ejection fraction (%)26.5 (21.4, 35.8)44.0 (30.0, 49.0)< .00138.7 (25.6, 47.0)
Hemoglobin (g/L)140.5 (135.0, 149.5)141.0 (129.0, 150.0).394141.0 (131.0, 150.0)
Creatinine (µmol/L)97.2 (79.6, 132.6)96.8 (81.3, 122.9).48897.2 (79.6, 123.8)
AHI (events/h)48.4 (38.8, 60.2)43.0 (31.8, 58.3).14044.6 (32.7, 59.4)
CAI (events/h)26.7 (16.6, 37.6)23.5 (14.2, 40.0).75224.2 (15.0, 39.9)
OAI (events/h)2.3 (0.7, 4.7)1.6 (0.4, 3.4).0871.7 (0.4, 3.7)
MAI (events/h)1.6 (0.4, 4.9)0.9 (0.2, 3.4).0461.1 (0.3, 3.9)
HI (events/h)10.7 (6.0, 20.7)11.6 (3.5, 19.8).39011.5 (3.7, 19.9)

Data presented as median (Q1, Q3) for continuous variables and percent (n) for categorical variables. a Nominal two-sided P from analysis of variance for continuous and Fisher exact test for categorical variables. AHI = apnea-hypopnea index, ALT = alanine aminotransferase, AST = aspartate aminotransferase, ASV = adaptive servoventilation, CAI = central apnea index, COPD = chronic obstructive pulmonary disease, HI = hypopnea index, MAI = mixed apnea index, NYHA = New York Heart Association, OAI = obstructive apnea index, PAV = positive airway ventilation, REM = rapid eye movement, SBP = systolic blood pressure, SDB = sleep-disordered breathing.

Device Implantation

Implant of the remedē System device was successful in 50 patients (88%) in the pilot study, including 44 on first attempt and all 6 who underwent a second attempt, and 147 (97%) in the pivotal trial on first attempt for a total of 197 of 208 patients (95%). Overall, 10 patients (5%) did not receive an implant because of anatomical issues preventing lead placement. The pooled mean time for successful implantation was 2.8 hours (standard deviation 0.9), with a mean of 3.0 hours in pilot and of 2.7 hours in pivotal. Pooled fluoroscopy time was 43.0 minutes, with a mean of 44.3 minutes in pilot and 42.6 minutes in pivotal. After the initial implant, 12 of the 50 patients (24%) of the pilot study underwent lead modification or explant in the first year due to issues such as lead dislodgment (stimulation lead pulled out of the target vessel and required the lead to be repositioned or replaced in order to deliver therapy) or lead displacement (stimulation lead remained in the target vessel but electrode position did not allow for effective therapy delivery). These lead complications were reduced during the pivotal trial (5/147; 3%) primarily due to new lead models being released prior to the pivotal trial. Over the course of the 2 clinical studies 17 (9%) of 197 patients underwent lead modification or explant in the first year due to lead issues such as lead dislodgment or lead displacement (12 pilot, 5 pivotal); of the 13 undergoing modification, 12 were successful.

Device Safety

There were 33 safety events in 30 patients (14%) through 12 months (Table 2). In the pivotal study fewer patients experienced safety events (n = 13, 9%) when compared to the pilot study (n = 17, 30%). Most pooled events were lead related (n = 18) but most of these events (n = 14) occurred during the pilot study prior to the new leads being available (right brachiocephalic vein lead for certain anatomic features). Ninety-three of the 208 patients had an existing cardiac implanted electronic device and 3 received one during the first year of follow-up. Among these patients, there was one reported serious device-device interaction resulting in an implantable cardioverter-defibrillator shock (pivotal) through 12 months. The interaction testing protocol was adjusted to be more robust based on this event.

Table 2 Summary of safety endpoints: related serious adverse events through 12 months.

ClassEventPilot (n = 57)Pivotal (n = 151)
Events% (n)Events% (n)
Any eventAny event2030 (17)139 (13)
Investigational device systemConcomitant device interaction00 (0)11 (1)
Device explantation12 (1)00 (0)
Extrarespiratory stimulation a24 (2)11 (1)
Inadequate lead position12 (1)00 (0)
Lead component failure12 (1)11 (1)
Lead dislodgement b1119 (11)21 (2)
Lead displacement c00 (0)11 (1)
Investigational device implant procedureImpending pocket erosion12 (1)21 (2)
Implant site hematoma12 (1)11 (1)
Implant site infection00 (0)21 (2)
OtherElevated transaminase00 (0)11 (1)
Headache12 (1)00 (0)
Noncardiac chest pain12 (1)11 (1)

a Discomfort of feeling delivered therapy in areas remote from the diaphragm. b When the stimulation lead pulled out of the target vessel and required the lead to be repositioned or replaced in order to deliver therapy. c When the lead remained in the target vessel but electrode position did not allow for effective therapy delivery.

Device Efficacy

After 6 months of therapy, the AHI in the pooled population was reduced from median (25th and 75th percentiles) 46.5 (34.1, 59.8) to 20.3 (9.8, 34.4) events/h. This represented a median reduction of 22.6 (8.4, 38.6) events/h (median 58% reduction from baseline, P < .001). This was driven by a median reduction in the CAI of 19.4 (9.9, 34.6) events/h (median 92% reduction from baseline, P < .001), resulting in a median CAI of 1.5 (0.2, 5.9) events/h at 6 months. (Figure 3, Table 3, and Figure S2, Figure S3, Table S3 and Table S4 in the supplemental material). Improvement in key sleep apnea indices (AHI, CAI) were maintained through 12-month follow-up. The pooled data indicated that a significant increase in REM sleep, improvement in arousals, and oxygenation (ODI4 and time with oxygen saturation < 90%) occurred at 6 months and were maintained at 12 months (Table 3).

Figure 3: Pooled polysomnography results from baseline to 6 and 12 months.

AHI = apnea-hypopnea index, ArI = arousal index, CAI = central apnea index, ODI4 = oxygen desaturation index of 4% or more.

Table 3 Summary of sleep metrics by months of active therapy in a pooled cohort.

EndpointBaseline6 Months of Therapy12 Months of Therapy
ResultChange From BaselinePResultChange From BaselineP
AHI (events/h)46.5 [34.1, 59.8] (203)20.3 [9.8, 34.3] (170)−22.6 [−38.6, −8.4] (170)< .001 a19.1 [10.1, 34.6] (156)−21.8 [−39.4, −7.8] (156)< .001 a
CAI (events/h)23.4 [13.9, 38.0] (203)1.5 [0.2, 5.9] (170)−19.4 [−34.6, −9.9] (170)< .001 a1.4 [0.3, 5.4] (156)−20.8 [−34.6, −12.2] (156)< .001 a
OAI (events/h)1.7 [0.5, 3.7] (203)2.3 [0.7, 7.7] (170)0.6 [−0.9, 4.5] (170)< .001 b3.0 [0.9, 7.1] (156)0.9 [−1.1, 5.1] (156)< .001 b
MAI (events/h)1.5 [0.3, 4.9] (203)0.0 [0.0, 0.5] (170)−0.9 [−3.5, 0.0] (170)< .001 b0.0 [0.0, 0.5] (156)−1.3 [−4.1, −0.1] (156)< .001 b
HI (events/h)11.6 [3.5, 19.8] (203)9.5 [3.9, 18.5] (170)−0.1 [−8.3, 4.9] (170).330 b10.0 [4.9, 19.2] (156)0.3 [−6.7, 7.0] (156).752 b
ODI4 (events/h)41.2 [25.7, 56.9] (203)18.8 [8.2, 33.0] (170)−18.8 [−33.4, −5.0] (170)< .001 a17.7 [8.8, 33.2] (156)−18.8 [−32.0, −5.8] (156)< .001 a
AI (events/h)39.5 [26.0, 54.4] (203)20.6 [13.9, 31.2] (170)−14.5 [−27.6, −3.3] (170)< .001 a22.9 [15.9, 34.9] (156)−12.1 [−28.5, −0.6] (156)< .001 a
O2 sat. < 90% (minutes)27.9 [6.5, 64.1] (203)13.1 [2.0, 40.1] (170)−8.0 [−31.1, 3.0] (170)< .001 b10.6 [2.2, 28.1] (156)−10.0 [−35.5, 1.4] (156)< .001 b
O2 sat. < 90% (%)8.4 [2.3, 20.4] (202)4.3 [0.7, 15.9] (170)−2.2 [−7.8, 1.9] (169)< .001 b3.9 [0.8, 11.9] (156)−2.6 [−9.1, 0.6] (155)< .001 b
REM (% of sleep)10.6 [5.4, 15.8] (203)15.1 [9.0, 19.9] (170)3.6 [−1.9, 8.5] (170)< .001 a15.5 [7.9, 21.5] (156)3.7 [−2.4, 9.5] (156)< .001 a
ESS (points)9.0 [5.0, 13.0] (203)6.0 [4.0, 9.0] (176)−2.0 [−6.0, 0.0] (176)< .001 b6.0 [4.0, 9.0] (153)−3.0 [−7.0, 0.0] (153)< .001 b

Data presented as median [Q1, Q3] (n). a Nominal two-sided P from paired t test for change from baseline to visit. b Nominal two-sided P from Wilcoxon signed-rank test for change from baseline to visit. AHI = apnea-hypopnea index, CAI, central apnea index, ESS = Epworth Sleepiness Scale, HI = hypopnea index, MAI = mixed apnea index, O2 sat. = oxygen saturation, OAI = obstructive apnea index, ODI4 = oxygen desaturation index of 4% or more, REM = rapid eye movement.

Patient-Reported Outcomes

Patients experienced improvement in daytime sleepiness as measured by a reduction in the Epworth Sleepiness Scale score at 6 months (from median [IQR] 9.0 [5.0, 13.0] to 6.0 [4.0, 9.0]) and 12 months (6.0 [4.0, 9.0]) (all P < .001). Additional functional assessment with a Patient Global Assessment tool performed by all patients and NYHA classification symptoms and Minnesota Living With Heart Failure Questionnaire completed by patients with baseline HF (Figure 4) indicated a significant improvement in symptoms and functional status at 6 months which was sustained at 12 months. At 12 months of active therapy, survival was 96% overall (Figure S4 in the supplemental material).

Figure 4: Patient-reported outcomes.

(A) Patient Global Assessment data, presented as change from baseline. (B) New York Heart Association classification assessment in patients with baseline heart failure. (C) Minnesota Living With Heart Failure Questionnaire medians and interquartile range by visit (pooled studies).

Cardiovascular Assessments

Use of the remedē System had no significant effects on clinical parameters such as resting blood pressure, heart rate, and weight (Table 3). Echocardiographic analysis of patients with heart failure and reduced ejection fraction (LVEF ≤ 45%) indicated a small but statistically significant improvement in LVEF from baseline, median (IQR) 27.0% (23.3, 36.0) to 31.1% (24.0, 41.5) at 12 months (P = .003), with parallel decrease in left ventricular volumes (Table 3).

Sensitivity Analyses

Sensitivity analysis using repeated measures identified potential study by visit interactions (P < .1) for two of the secondary endpoints used in the pivotal trial (arousal index and percent of sleep in REM), indicating a potentially different direction or magnitude of results among the studies and visits that could suggest pooling these endpoints might not be appropriate. However, the primary differences driving these two findings were the magnitude of the effect at different visits. For example, the arousal index improved by 11.5 and 3.3 events/h at 6 and 12 months, respectively, in the pilot study compared to improvements of 17.8 and 17.7 events/h at 6 and 12 months in the pivotal study. The detailed analysis is presented in Table S4.


This analysis represents the largest experience with implantable transvenous PNS for the treatment of CSA. There are four major findings from these pooled data from the remedē System pilot and pivotal studies. First, the therapy was safe and was characterized by a 95% device implant success rate and improvements in implant efficiency over time. Second, PNS resulted in a substantial decrease in CSA burden and concordant improvements in sleep measures and QOL through 12 months. Third, in patients with HF, HF symptoms improved at 6 months and in those with LVEF ≤ 45%, cardiac function improved at 6 months. Importantly, all benefits of phrenic stimulation observed at 6 months were sustained at 12 months.

CSA is commonly encountered in clinical practice, frequently associated with symptomatic HF, and associated with poor outcomes. Currently, PAP is the mainstay treatment for CSA. Trials of PAP for CSA have yielded conflicting results.20 In fact, the recently published Adaptive Servo-Ventilation for Central Sleep Apnea in Systolic Heart Failure Trial (SERVE-HF) raised concerns of increased mortality with the use of adaptive servoventilation, a form of positive pressure ventilation, for the treatment of patients with HF with reduced ejection fraction and predominant CSA.12,21

The pooled data from the pilot and pivotal trials confirm that PNS therapy is well-tolerated though 12 months. The proposed benefit of PNS when compared to PAP is that it induces a physiologic breathing pattern through diaphragmatic stimulation with resultant negative intrathoracic pressures. As a result, the physiologic effects of diaphragmatic stimulation do not exert the same adverse hemodynamic impact on central hemodynamics as those seen with PAP ventilation (eg, increased intrathoracic pressure affecting right and left ventricular preload and afterload).8,12,2124 The recent approval of the PNS system in Europe and the United States now provides a novel device-based therapy for patients with CSA. The pooled analysis of the two clinical studies provides a unique insight in the sustained safety and effectiveness of the therapy across different trials, increased number of implanters and with a long-term follow-up. As expected with novel device therapies, the implant success and procedural complication rate improved from the pilot study to the pivotal trial. Several factors may have influenced the improvements in implant success rate, including increased operator experience, better leads, and modified implantation techniques. Notably, the rate of complications and need for modifications was comparable to other transvenous technologies such as cardiac resynchronization therapy at a similar time point in their clinical developments.25,26

The median 58% improvement in AHI was paralleled by significant reductions in the burden of CSA and percentage of time spent with oxygen saturation < 90%, a measure that is highly associated with long-term outcomes.27 The improvement in sleep quality and maladaptive sympathetic activation, as suggested by a lower arousal index, is unique to PNS and not previously observed in large trials of PAP therapy for the treatment of CSA.11,12 Decreased frequency of arousals and hypoxic episodes could alleviate the accompanying sympathetic surges.8,12,28 The consistency and degree of improvement in both sleep and QOL parameters suggests that the effect of the physiologic neurostimulation therapy is clinically relevant, affecting a number of patient-centered outcomes. Notably, although central apnea events are reduced to a minimal level while on therapy (CAI of 1.5 events/h [25th and 75th percentiles; 0.2 and 5.9, respectively] at 6 months), residual episodes are likely the result of either concomitant obstructive episodes that were not affected by PNS or centrally mediated events (apnea or hypopnea) that occurred during automated short suspensions of therapy that occur during light stages of sleep (ie, following a roll or bathroom break) or while therapy is automatically ramping up to the programmed effective therapy level at the start of the night or after a suspension. This device automaticity is intended to allow the patient to fall asleep and remain asleep, maximizing the duration of the night therapy is on.

Importantly, the positive effects extended to the sickest subgroup of patients with CSA, specifically those with HF. In the pooled cohort analysis, most patients (68%) had a concomitant diagnosis of HF at baseline. Nevertheless, a recent post hoc analysis from the pivotal trial indicates that the efficacy and safety of PNS are similar in patients with CSA who have versus those who do not have concomitant HF.29 The findings of favorable effects on cardiac function and HF-specific QOL measured in each study were confirmed and expanded by the results of the current analysis. Additionally, the robustness of the effects of PNS on CSA is supported by the consistency of the results across different time periods and study sites.

Clinical Implications

The significant and long-term reduction of key determinants of CSA severity, such as AHI, mixed apnea index, CAI, and 4%ODI, established the feasibility and therapeutic efficacy of unilateral PNS in CSA. As an example, the degree of AHI and the degree of reduction in AHI has been associated with improved outcomes in patients with either obstructive or CSA.3,3032 Whether or not the reduction in key sleep parameters, symptoms, and cardiac function by the remedē System can positively affect cardiovascular outcomes remains to be determined. Thus, future trials should test the physiologic effects on the cardiovascular system and examine the hypothesis that PNS can improve cardiovascular outcomes such as HF-related hospitalizations and mortality in patients with CSA.


Similarities in inclusion and exclusion criteria, trial visit schedule, and endpoints allowed pooling the two original pilot and pivotal studies, yet a few study/visit interactions were present. However, the differences between the pilot and pivotal studies were driven by the magnitude of change at visits for the respiratory and sleep metrics as opposed to differences in the direction of response. The current pooled cohort analysis was limited by the open-label design of the studies. Self-reported measures of health status and symptoms could lead to bias. However, objective measures of CSA and cardiac function were evaluated by core laboratories blinded to therapy being active or not and additionally to treatment assignment in the pivotal study. Notably, because of the higher prevalence of CSA in male patients, men predominated the current cohort, thus limiting any statements regarding the effects of PNS in women with CSA. Furthermore, the predominance of Caucasians race limits the ability to extrapolate the effects of PNS to other racial groups. Finally, although no adverse cardiovascular effects on clinical parameters were seen and, in fact, an improvement in echocardiographic measures of cardiac function was observed, future mechanistic work should be considered to further elucidate the effects of PNS on cardiac functional and hemodynamic features.


To date the combined evidence for transvenous PNS for CSA indicates a significant reduction in the severity of CSA, with improvements in the arousal index, self-reported daytime sleepiness, REM sleep, cardiac function, symptoms, and QOL measures up to 12 months. This new neurostimulation approach is a safe and effective treatment for CSA that is independent of patient adherence.


All authors have seen and approved the manuscript. The remedē pilot and pivotal clinical studies were sponsored by Respicardia Inc, Minnetonka, MN, USA. No financial support was provided for the preparation of this manuscript. Patient level data and statistics were handled by Mr. McKane, who is an employee of Respicardia Inc, Minnetonka, MN, USA. Dr. Fudim reports support from an American Heart Association Grant, 17MCPRP33460225 and NIH T32 grant 5T32HL007101; and consulting fees/honoraria from Coridea, AxonTherapies. Dr. Costanzo reports consulting for Respicardia. Dr. Pokorney reports research support from Medtronic, Boston Scientific, Gilead, Pfizer, Bristol-Myers Squibb, and Janssen Pharmaceuticals; and consulting/advisory board support from Boston Scientific, Medtronic, Pfizer, Bristol-Myers Squibb, and Janssen Pharmaceuticals. Dr. Abraham reports consulting for Respicardia. Dr. Piccini reports support from grants for clinical research from Abbott, ARCA biopharma, Boston Scientific, Gilead, Janssen Pharmaceuticals, and NHLBI and consulting for Abbott, Allergan, ARCA Biopharma, Bayer, Biotronik, GSK, Johnson & Johnson, Medtronic, Motif Bio, Sanofi, and Phillips. The other authors report no conflicts of interest.



apnea-hypopnea index


central apnea index


central sleep apnea


phrenic nerve stimulation


quality of life


rapid eye movement


sleep-disordered breathing


  • 1. Lanfranchi PA, Somers VK, Braghiroli A, Corra U, Eleuteri E, Giannuzzi P. Central sleep apnea in left ventricular dysfunction: prevalence and implications for arrhythmic risk. Circulation. 2003;107(5):727–732.

    CrossrefGoogle Scholar
  • 2. Oldenburg O, Lamp B, Faber L, Teschler H, Horstkotte D, Topfer V. Sleep-disordered breathing in patients with symptomatic heart failure: a contemporary study of prevalence in and characteristics of 700 patients. Eur J Heart Fail. 2007;9(3):251–257.

    CrossrefGoogle Scholar
  • 3. Javaheri S, Shukla R, Zeigler H, Wexler L. Central sleep apnea, right ventricular dysfunction, and low diastolic blood pressure are predictors of mortality in systolic heart failure. J Am Coll Cardiol. 2007;49(20):2028–2034.

    CrossrefGoogle Scholar
  • 4. Hanly PJ, Zuberi-Khokhar NS. Increased mortality associated with Cheyne-Stokes respiration in patients with congestive heart failure. Am J Respir Crit Care Med. 1996;153(1):272–276.

    CrossrefGoogle Scholar
  • 5. Somers VK, White DP, Amin R, et al.. Sleep apnea and cardiovascular disease: an American Heart Association/American College of Cardiology Foundation scientific statement from the American Heart Association Council for High Blood Pressure Research Professional Education Committee, Council on Clinical Cardiology, Stroke Council, and Council on Cardiovascular Nursing. In collaboration with the National Heart, Lung, and Blood Institute National Center on Sleep Disorders Research (National Institutes of Health). Circulation. 2008;118(10):1080–1111.

    CrossrefGoogle Scholar
  • 6. Aurora RN, Chowdhuri S, Ramar K, et al.. The treatment of central sleep apnea syndromes in adults: practice parameters with an evidence-based literature review and meta-analyses. Sleep. 2012;35(1):17–40.

    CrossrefGoogle Scholar
  • 7. Germany R. Non-mask-based therapies for central sleep apnea in patients with heart failure. Sleep Med Clin. 2017;12(2):255–264.

    CrossrefGoogle Scholar
  • 8. Costanzo MR, Khayat R, Ponikowski P, et al.. Mechanisms and clinical consequences of untreated central sleep apnea in heart failure. J Am Coll Cardiol. 2015;65(1):72–84.

    CrossrefGoogle Scholar
  • 9. Khayat R, Abraham W, Patt B, et al.. Central sleep apnea is a predictor of cardiac readmission in hospitalized patients with systolic heart failure. J Card Fail. 2012;18(7):534–540.

    CrossrefGoogle Scholar
  • 10. Khayat R, Jarjoura D, Porter K, et al.. Sleep disordered breathing and post-discharge mortality in patients with acute heart failure. Eur Heart J. 2015;36(23):1463–1469.

    CrossrefGoogle Scholar
  • 11. Bradley TD, Logan AG, Kimoff RJ, et al.. Continuous positive airway pressure for central sleep apnea and heart failure. N Engl J Med. 2005;353(19):2025–2033.

    CrossrefGoogle Scholar
  • 12. Cowie MR, Woehrle H, Wegscheider K, et al.. Adaptive servo-ventilation for central sleep apnea in systolic heart failure. N Engl J Med. 2015;373(12):1095–1105.

    CrossrefGoogle Scholar
  • 13. Zhang XL, Ding N, Wang H, et al.. Transvenous phrenic nerve stimulation in patients with Cheyne-Stokes respiration and congestive heart failure: a safety and proof-of-concept study. Chest. 2012;142(4):927–934.

    CrossrefGoogle Scholar
  • 14. Costanzo MR, Augostini R, Goldberg LR, Ponikowski P, Stellbrink C, Javaheri S. Design of the remede System pivotal trial: a prospective, randomized study in the use of respiratory rhythm management to treat central sleep apnea. J Card Fail. 2015;21(11):892–902.

    CrossrefGoogle Scholar
  • 15. Abraham WT, Jagielski D, Oldenburg O, et al.. Phrenic nerve stimulation for the treatment of central sleep apnea. JACC Heart Fail. 2015;3(5):360–369.

    CrossrefGoogle Scholar
  • 16. Costanzo MR, Ponikowski P, Javaheri S, et al.. Transvenous neurostimulation for central sleep apnoea: a randomised controlled trial. Lancet. 2016;388(10048):974–982.

    CrossrefGoogle Scholar
  • 17. Berry RB, Brooks R, Gamaldo CE, et al.; for the American Academy of Sleep Medicine. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications. Version 2.0.. Darien, IL: American Academy of Sleep Medicine; 2012.

    Google Scholar
  • 18. Jagielski D, Ponikowski P, Augostini R, Kolodziej A, Khayat R, Abraham WT. Transvenous stimulation of the phrenic nerve for the treatment of central sleep apnoea: 12 months’ experience with the remede((R)) System. Eur J Heart Fail. 2016;18(11):1386–1393.

    CrossrefGoogle Scholar
  • 19. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540–545.

    CrossrefGoogle Scholar
  • 20. Yu J, Zhou Z, McEvoy RD, et al.. Association of positive airway pressure with cardiovascular events and death in adults with sleep apnea: a systematic review and meta-analysis. JAMA. 2017;318(2):156–166.

    CrossrefGoogle Scholar
  • 21. Cowie MR, Wegscheider K, Teschler H. Adaptive servo-ventilation for central sleep apnea in heart failure. N Engl J Med. 2016;374(7):690–691.

    Google Scholar
  • 22. Le Pimpec-Barthes F, Gonzalez-Bermejo J, Hubsch JP, et al.. Intrathoracic phrenic pacing: a 10-year experience in France. J Thorac Cardiovasc Surg. 2011;142(2):378–383.

    CrossrefGoogle Scholar
  • 23. Combes N, Jaffuel D, Cayla G, et al.. Pressure-dependent hemodynamic effect of continuous positive airway pressure in severe chronic heart failure: a case series. Int J Cardiol. 2014;171(3):e104–e105.

    CrossrefGoogle Scholar
  • 24. Liston R, Deegan PC, McCreery C, Costello R, Maurer B, McNicholas WT. Haemodynamic effects of nasal continuous positive airway pressure in severe congestive heart failure. Eur Respir J. 1995;8(3):430–435.

    CrossrefGoogle Scholar
  • 25. Linde C, Abraham WT, Gold MR, et al.. Randomized trial of cardiac resynchronization in mildly symptomatic heart failure patients and in asymptomatic patients with left ventricular dysfunction and previous heart failure symptoms. J Am Coll Cardiol. 2008;52(23):1834–1843.

    CrossrefGoogle Scholar
  • 26. Alonso C, Leclercq C, d’Allonnes FR, et al.. Six year experience of transvenous left ventricular lead implantation for permanent biventricular pacing in patients with advanced heart failure: technical aspects. Heart. 2001;86(4):405–410.

    CrossrefGoogle Scholar
  • 27. Oldenburg O, Wellmann B, Buchholz A, et al.. Nocturnal hypoxaemia is associated with increased mortality in stable heart failure patients. Eur Heart J. 2016;37(21):1695–1703.

    CrossrefGoogle Scholar
  • 28. Javaheri S, Parker TJ, Liming JD, et al.. Sleep apnea in 81 ambulatory male patients with stable heart failure. Types and their prevalences, consequences, and presentations. Circulation. 1998;97(21):2154–2159.

    CrossrefGoogle Scholar
  • 29. Costanzo MR, Ponikowski P, Coats A, et al.. Phrenic nerve stimulation to treat patients with central sleep apnoea and heart failure. Eur J Heart Fail. 2018;20(12):1746–1754.

    CrossrefGoogle Scholar
  • 30. Jilek C, Krenn M, Sebah D, et al.. Prognostic impact of sleep disordered breathing and its treatment in heart failure: an observational study. Eur J Heart Fail. 2011;13(1):68–75.

    CrossrefGoogle Scholar
  • 31. Hastings PC, Vazir A, Meadows GE, et al.. Adaptive servo-ventilation in heart failure patients with sleep apnea: a real world study. Int J Cardiol. 2010;139(1):17–24.

    CrossrefGoogle Scholar
  • 32. Arzt M, Floras JS, Logan AG, et al.. Suppression of central sleep apnea by continuous positive airway pressure and transplant-free survival in heart failure: a post hoc analysis of the Canadian Continuous Positive Airway Pressure for Patients with Central Sleep Apnea and Heart Failure Trial (CANPAP). Circulation. 2007;115(25):3173–3180.

    CrossrefGoogle Scholar