Abstract
Diaphragm pacing with intramuscular electrodes causes pain in central hypoventilation patients without spinal lesions http://ow.ly/IItjY
To the Editor:
Diaphragm pacing is an alternative to mechanical ventilation in patients with high cervical cord damage (>C4 dermatome). It brings clinical benefits and reduces health costs [1–3]. It is also indicated in certain cases of central hypoventilation [4]. Two types of diaphragm pacing devices are commercially available. With intrathoracic phrenic stimulation, electrodes are implanted around the phrenic nerves in the thorax (Avery Biomedical, Commack, NY, USA; and Atrotech, Tampere, Finland). Nerve dissection during a video-assisted mini-thoracotomy allows intimate contact of the stimulating electrodes with the nerve [5]. Stimulating currents typically range between 0.3–3.5 mA [6]. With intradiaphragmatic stimulation, hook wire electrodes are laparoscopically implanted in the diaphragm in the vicinity of the phrenic nerve termination (Synapse Biomedical, USA) [in 5]. Stimulating currents typically range between 5–20 mA. In quadriplegic patients, intradiaphragmatic phrenic stimulation is effective [7] and has been advocated as easier to implement and less expensive than its intrathoracic counterpart [5]. In patients who retain spontaneous diaphragm activity, this approach is appealing because it is devoid of the theoretical risk of procedure-induced phrenic nerve damage associated with the intrathoracic route. For these reasons, we implanted intradiaphragmatic phrenic stimulators in four hypoventilation patients after the device was authorised in France in 2010. In France, both intrathoracic and intradiaphragmatic diaphragm pacing are authorised and reimbursed by social security in quadriplegia and central hypoventilation and are managed at a single multidisciplinary centre nationwide.
This report is motivated by the fact that, in these four cases, diaphragm pacing efficiency was compromised and clinical management complicated by phrenic stimulation associated pain. We did not observe this in comparable patients implanted intrathoracically.
We describe nine patients (table 1) who all had central hypoventilation, documented diaphragm responses to phrenic stimulation and preserved sensitivity on clinical examination (present and symmetrical pin-prick and light-touch sensations on routine systematic neurological examination). They reported normal pain perception in daily life. Phrenic stimulator implantations were performed over a 15-year period. Patients 1–5 were implanted intrathoracically (single surgeon, group 1) and patients 6–9 were implanted intradiaphragmatically (single surgeon, group 2). Patients 1–3 were implanted first, then patients 6–9 and then patients 4–5. Patients 1–3 were implanted in the context of an externally approved research [2]. The other patients were implanted on clinical indications. They gave consent to the anonymous use of their clinical data, which was approved by the Institutional Review Board of the Société de Pneumologie de Langue Française (decision #2014-048). The reversal of hypoventilation-related pulmonary hypertension by intradiaphragmatic phrenic stimulation in patient 6 has been described elsewhere [8].
Phrenic stimulation-induced pain was defined as pain: 1) reported spontaneously or in response to oriented questioning, 2) appearing upon start of pacing and disappearing more or less rapidly upon its discontinuation, 3) involving a pathophysiologically logical territory (upper abdominal quadrants, inferior thoracic regions, neck/shoulder), and 4) requiring both a reduction in stimulation intensity and analgesic medications to pursue pacing. Full ventilatory autonomy (weaning from mechanical ventilation) was defined as the possibility for the patient to remain on diaphragm pacing 24 h a day, irrespective of the actual use of the device and of arterial carbon dioxide tension (PaCO2). Some patients reached ventilatory autonomy a few days after implantation. In others, a weaning protocol similar to that used in quadriplegia was followed (incremental daily stimulation sessions guided on tidal volume maintenance and clinical tolerance) for 2–9 weeks. For the purpose of this particular report, the ventilatory autonomy outcome was assessed 1 year after implantation. Persistent hypoventilation was defined as hypercapnia (PaCO2 ≥45 mmHg) under pacing. Full correction of pulmonary hypertension was defined as the return of echocardiographically measured systolic arterial pressure <30 mmHg. Outcomes were compared using 2×2 contingency tables and Fischer's exact test with p<0.05 considered significant.
Phrenic stimulation-induced pain was never noted in group 1, while it was always present in group 2 (p=0.0079). It consisted in unilateral or bilateral neck and shoulder pain (C3–C5 dermatomes). Pain was generally described as an aching and/or burning sensation that started immediately or quasi so after switching the stimulator on and could last several minutes after switching it off. It was not associated with allodynia or hyperalgesia. Before any intervention, visual analogue scale assessment ranged from 20–70% of the full scale, depending on patients. Ventilatory autonomy was consistently achieved in group 1 and in two Group 2 patients (p=0.166). Persistent hypoventilation was not noted in group 1 but present in three group 2 patients (p=0.047). Correction of pulmonary hypertension was achieved irrespectively of the stimulation technique (one case in each group). One patient (patient 6) who was prescribed pregabalin to improve pacing tolerance developed cytolytic hepatitis requiring a change in treatment.
Although the observations were gathered over a long period of time, the pre-implantation procedures, the surgical procedures and the follow-up procedures were invariant. All the patients were tested by the same investigators, the stimulators were implanted by a single surgeon for each technique, and the follow-up was standardised at a single centre. In addition, the intradiaphragmatic series was “bracketed” by the intrathoracic series. We, therefore, believe that the length of the observation span (that is easily explained by the unique nature of the cohort; to our knowledge, there is no description of a similar case series of adult hypoventilation patients in the literature) is not a significant source of bias.
The phrenic nerve is a mixed nerve. It carries afferents from the subdiaphragmatic peritoneum (liver and spleen), pericardium, lower regions of the pleura, and the diaphragm [9]. Clinically, phrenic afferents irritation translates in referred neck and shoulder pain (C3–C5 sensory territory, Kehr's sign) [10]. Our observations are compatible with this mechanism. Phrenic nerve afferents comprise diaphragmatic C-fibres [11]. These small unmyelinated fibres have a high excitation threshold; they are not likely to be depolarised by the low-intensity currents used for intrathoracic phrenic stimulation (maximum 2.2 mA in table 1) that are permitted by the close electrode-nerve proximity. We hypothesise that, in contrast, the higher stimulation intensities (table 1) required to achieve intradiaphragmatic phrenic stimulation, because of a greater electrode-nerve distance, were sufficient to depolarise C-fibres in the vicinity of the electrodes and induce pain. In line with this, analgesic drugs known to be efficient on neuropathic pain (pregabalin, gabapentin and duloxetine) proved useful in our patients. Reducing stimulation intensities also proved useful but compromised pacing efficiency; even though the frequency of ventilatory autonomy was not statistically different between groups, persistent hypoventilation was significantly more frequent in group 2. Diaphragm pacing associated pain has not been reported with intrathoracic phrenic stimulation except in cases of device dysfunction [12]. Intradiaphragmatic phrenic nerve stimulation-related “discomfort” [13] or overt pain [14], have been described in clinical trials of diaphragm pacing in amyotrophic lateral sclerosis. In this context, phrenic stimulation does not aim at achieving ventilation but at diaphragm conditioning; it is delivered at lower intensities that can be further decreased to control pain. Yet the occurrence of pain in this setting confirms the reality of the issue. We also observed pain during intradiaphragmatic phrenic stimulation in two quadriplegic patients with incomplete spinal cord lesions. Of note, blunted pain perception has been described in congenital central hypoventilation syndrome (CCHS) children [4, 15] and could have contributed to the lack of pain reported in the two CCHS patients in group 1. Yet the other patients in this group did not have pain perception issues and the CCHS patient in group 2 did experience severe pain.
In conclusion, our observations should be borne in mind when choosing a diaphragm pacing technique in patients with functional sensory pathways. The advantages of intradiaphragmatic phrenic pacing could indeed then be offset by tolerance issues. Future technical developments of intradiaphragmatic phrenic stimulation should focus on stimulation protocols preserving stimulation of motor fibres while avoiding C-fibres stimulation. Unreported preliminary data suggest that modified pulse modulation schemes could achieve this, but this will need to be specifically studied. Of importance, future other mini-invasive diaphragm pacing techniques, such as transvenous phrenic stimulation [16], should be scrutinised for tolerance outside the particular quadriplegia context and particularly if diaphragm pacing indications widen, e.g. as an adjunct to mechanical ventilation in intensive care unit patients [17].
Acknowledgements
The authors are grateful to Paul Robinson for editing English style and grammar.
Footnotes
Support statement: This work was supported by the programme “Investissement d'Avenir ANR-10-AIHU 06” of the French Government and the non-profit research association Association pour le Développement et l’Organisation de la Recherche en Pneumologie et sur le Sommeil (ADOREPS), Paris, France. Some of the patients described were included in a study of diaphragm pacing supported by grant DRC98075 from the Programme Hospitalier de Recherche Clinique National of the French Ministry of Health and of which the sponsor was the Direction de la Recherche Clinique, Assistance Publique-Hôpitaux de Paris, Paris, France. Funding information for this article has been deposited with FundRef
Conflict of interest: Disclosures can be found alongside the online version of this article at erj.ersjournals.com
- Received November 29, 2014.
- Accepted January 3, 2015.
- Copyright ©ERS 2015