There is a clinically significant association between central sleep apnea (CSA) and
heart failure with reduced ejection fraction (HFrEF).
The presence of CSA has been associated with an increased risk of decompensated heart
failure and death
and the prevalence of CSA increases in parallel with the severity of HFrEF. Phrenic
nerve stimulation (PNS) using the remedē system (Respicardia, Minnetonka, MN) is a
new therapeutic modality to treat CSA that involves placing transvenous leads to stimulate
the phrenic nerve and activate the diaphragm. Since many patients with severe HFrEF
have implanted defibrillators, a clinical scenario where concomitant CSA and defibrillator
therapy are necessary is likely to occur. However, data are lacking concerning potential
interaction between these devices. We present a case report of a patient with severe
ischemic cardiomyopathy with a pre-existing subcutaneous implantable cardioverter-defibrillator
(S-ICD) who was later diagnosed with severe CSA and underwent successful remedē system
A 54-year-old man with poorly controlled type 1 diabetes mellitus, end-stage kidney
disease on hemodialysis, paroxysmal atrial fibrillation, and ischemic cardiomyopathy
with HFrEF (left ventricular ejection fraction 30%) underwent a primary prevention
Emblem S-ICD insertion (Boston Scientific, St. Paul, MN) in December 2015. The sensing
vector was secondary. Conversion testing of ventricular fibrillation was successfully
performed with 65 J shock and the shock impedance was measured as 46 ohms.
In November 2018, an in-laboratory polysomnogram was obtained to evaluate for obstructive
sleep apnea owing to complaints of daytime sleepiness, fatigue, and difficulty initiating
and maintaining sleep. At that time his body mass index was 20.4 kg/m2. His most recent
echocardiogram demonstrated further worsening of heart function with a left ventricle
ejection fraction of 14%. The polysomnogram revealed severe sleep-disordered breathing
with predominant CSA. The total apnea-hypopnea index was 61 events per hour (normal:
less than 5 events per hour; severe: more than 30 events per hour). The central apnea
index was 42 events per hour and the obstructive apnea-hypopnea index was 19 events
per hour. Sleep-disordered breathing was also accompanied by significant sleep hypoxemia.
Figure 1 illustrates a representative 2-minute window from the overnight polysomnogram
demonstrating central apneas. After failing continuous positive airway pressure (CPAP)
titration, the patient underwent bilevel positive airway pressure spontaneous-timed
(PAP-ST) titration in the sleep laboratory and was prescribed home PAP-ST therapy
with settings of 25/16 cm H2O with a backup respiratory rate of 16 breaths per minute
with an oronasal mask. Adequate adherence was shown during the first 30 days of therapy,
as obtained by downloading data from the patient’s device. The device-estimated residual
apnea-hypopnea index was 22 events per hour (residual central apnea index of 14 and
hypopnea index of 8). He did not feel any improvement in his symptoms despite adequate
adherence to bilevel PAP-ST therapy. Given the persistence of symptoms and persistent
CSA despite bilevel PAP-ST therapy, he was referred for phrenic nerve stimulator implantation.
Polysomnogram showing central apneas. Two-minute window from overnight polysomnogram
demonstrating frequent central apneas with Cheyne-Stokes respiration. There are 2
central apneas during the representative 2-minute window recorded during stage 2 non–rapid
eye movement (N2) sleep based on the 6 electroencephalography leads (F3-M2, C3-M2,
O1-M2, F4-M1, C4-M1, O2-M1) with the patient in the right lateral decubitus position.
Each central apnea lasts approximately 30 seconds and is characterized by a lack of
airflow on the nasal pressure transducer (NPT) and the oronasal thermistor (Therm),
no effort on the chest and abdominal respiratory inductance plethysmography belts,
followed by a 10% oxygen desaturation (in yellow). Each cycle, from the beginning
of an apnea to the end of the next apnea, lasts approximately 60 seconds and the microarousals
occur at the peak of ventilatory effort, consistent with central sleep apnea with
Cheyne-Stokes respiration pattern.
Implantation of the remedē system was performed in the electrophysiology laboratory
under moderate sedation via the right subclavian vein access. The S-ICD was temporarily
disabled to allow for electrocautery. The implantation technique of the remedē system
has been described previously
but in short, we inserted a multipolar pacing lead Respistim LQS model 4065 (Respicardia,
Minnetonka, MN,) into the left pericardiophrenic vein, which courses over the lateral
border of the epicardium and runs parallel with the left phrenic nerve. A second lead,
Medtronic Attain Ability model 4196 (Medtronic, Minneapolis, MN), was inserted into
a low right intercostal vein via the azygos vein. This lead is used for monitoring
transthoracic impedance, which is a surrogate for respiration. The generator was placed
in a pectoral pocket (Figure 2).
Chest radiograph showing implanted hardware. Anteroposterior (AP) chest radiograph
of right pectoral position of phrenic nerve stimulator with multipolar pacing lead
in left pericardiophrenic vein and sensing lead in low right intercostal vein via
the azygos vein. Subcutaneous implantable cardioverter-defibrillator (S-ICD) is in
its usual position on the left lateral thorax with corresponding suprasternal lead.
A = distal electrode; AL = alternate S-ICD vector; B = proximal electrode; PR = primary
S-ICD vector; SC = secondary S-ICD vector.
Because of this patient’s pre-existing S-ICD, the potential for device-device interaction
was systematically assessed. The S-ICD was re-enabled. With maximum Respistim LQS
bipolar pacing output (10 mA, 300 microseconds, 20 Hz) from all 4 pacing electrodes,
subcutaneous signals were recorded from all 3 S-ICD vectors. Pacing artifact was evident
in the primary and secondary vectors but absent in the alternate. While the pacing
artifact was visible, it was not sensed as noise. During the remedē impedance check,
4 0.5-mA pulses were delivered and the resultant artifact was recorded on the subcutaneous
signal. The sinus beat preceding the impedance check stimulus was labeled as noise
and the stimulus was sensed (Figure 3). The S-ICD sensing vector was programmed as
secondary. The leads were connected to the generator and the system was implanted
in a right pectoral pocket with the aid of a TYRX antimicrobial envelope (Medtronic,
St. Paul, MN). The remedē system was not activated, as is recommended, to allow for
the leads to settle in and stabilize. The patient was discharged the next day and
6 weeks later was brought back for remedē system activation. With S-ICD sensing programmed
to secondary vector, device-device interaction was assessed in the same fashion as
at the time of the initial implant. There was no noise detected with up-titration
of pacing output. With impedance testing, artifact similar to what was seen during
implant was present, but this finding was not sensed reproducibly. The remedē system
was activated. After 6 weeks of PNS, he is feeling better and S-ICD interrogation
has not revealed any shocks. A repeat polysomnogram is scheduled in 2 months.
Subcutaneous electrocardiography signals recorded from the subcutaneous implantable
cardioverter-defibrillator during phrenic nerve stimulation impedance testing. During
impedance testing by the remedē device (Respicardia, Minnetonka, MN), the stimulus
(red arrow) was recorded and labeled as a sensed event (S). The preceding sinus beat
(which is on time with the patient’s intrinsic rhythm) was labeled as noise (N). This
finding was reproducible during implant but not evident at the follow-up visit.
To our knowledge, this is the first report of concomitant S-ICD and phrenic nerve
stimulator implantation. The prevalence of sleep apnea, both obstructive and central,
is high among patients with congestive heart failure, approaching 50%–75% in some
Untreated sleep apnea is associated with high morbidity and mortality.
Obstructive sleep apnea is more commonly recognized and is usually treated with positive
pressure therapy such as CPAP. Prior to PNS, there were no effective treatments for
CSA, as CPAP does not effectively treat central apneas
and adaptive servoventilation has been associated with a higher mortality in patients
particularly in patients with an ejection fraction below 30%.
PNS with the remedē system has been shown to improve symptoms and reduce central apneas
and is currently indicated for patients diagnosed with moderate-to-severe CSA.
Both sensing and pacing leads of the remedē system are programmable. The quadripolar
left lead can be programmed in multipolar pacing configuration to deliver output between
0.1 and 10 mA at a pulse width of 60–300 microseconds. Most patients require up-titration
of PNS pacing output during follow-up, which is why it is important to test for interaction
at maximum output and with all available pacing options.
There are 3 sensing vectors available on the S-ICD that are used for detecting malignant
ventricular arrhythmias (primary, secondary, and alternate) (Figure 2). The S-ICD
algorithm automatically chooses one based on subcutaneous signal amplitude, morphology,
A different sensing vector can be selected by the implanting physician if desired.
Although an artifact was detected on primary and secondary vectors during pacing and
impedance testing of the remedē system, the alternate vector was spared, as it senses
signals from an area outside the PNS (ie, between distal and proximal lead electrodes;
Figure 2). While it may seem desirable to manually select the alternate vector in
these cases, it is often the one least chosen by the S-ICD algorithm owing to poor
Regardless, even with maximum pacing output from the PNS, there was no noise or pacing
artifact sensed. While artifact was recorded and intermittently sensed during PNS
impedance testing, this finding was not reproducible. Because the S-ICD has limited
storage capability, impedance testing artifact, even if it occurred during auto–impedance
testing by the remedē device, would not likely be saved and available for review.
Our case is the first example of concomitant PNS and S-ICD implantation. Given the
relatively high prevalence of CSA in the HFrEF population, situations where these
2 therapies could be utilized concurrently will become more commonplace. We did not
see any significant interactions, but our follow-up is short. More data are needed
before we can definitively say that concomitant PNS and S-ICD therapy is safe.
Key Teaching Points
The prevalence of central sleep apnea (CSA) is relatively high in patients with heart
failure with reduced ejection fraction and can lead to significant morbidity and mortality
if left untreated. Phrenic nerve stimulation (PNS) is a desirable option in patients
with systolic dysfunction, as continuous positive airway pressure is not effective
in treating CSA.
The artifact generated during impedance testing by the remedē device (Respicardia,
Minnetonka, MN) could theoretically interfere with subcutaneous implantable cardioverter-defibrillator
function but is not demonstrated in this case despite rigorous testing.
Because up-titration of pacing output and/or pacing pole reprogramming is often needed
after PNS implant, potential device-device interaction should be assessed at implant
and during follow-up using maximum pacing output in all available PNS pacing configurations.