Abstract
Chronic
aspiration following stroke is believed to lead to nearly 40,000 deaths from
pneumonia in the
Introduction
According to the National
Stroke Association, 550,000 Americans suffer a new or recurrent stroke each
year. Of these patients, 19% to 38%, or
about 150,000, will aspirate[1], which is defined as the taking of foreign material,
such as saliva and ingested food, into the lungs. Aspiration increases the chances of
contracting pneumonia by 20-fold and is believed responsible for approximately
40,000 deaths following a stroke every year in the Unites States alone [2].
Aspiration results from several factors, including insufficient
laryngeal elevation, poor coordination of the swallowing effort, and/or
insufficient closure of the vocal folds to protect the airway. If standard therapeutic treatments fail, they
patient may not be allowed to take food by mouth and enteral
means of nutrition are required. Current
options to protect the airway involve irreversible and destructive surgery of
the upper airway. These often reduce the
patient’s quality of life.
The authors propose that stimulation of the recurrent laryngeal nerve will adduct the vocal folds, closing the airway and reducing aspiration. The feasibility of this approach has been demonstrated in animal models [3, 4]. Here the authors present the first clinical trial demonstrating the feasibility of this approach. The two hypotheses of the study are 1) that stimulation of the vocal folds in humans will adduct the vocal cords to close the airway and 2) vocal fold adduction (VFAd) will reduce aspiration.
Methods
Patient Selection
The complete study will
enroll ten control subjects (receiving only standard treatment) and ten
experimental subjects (receiving the stimulation system), each group randomly
assigned. Eligible patients are at least
six months post-stroke and have already been treated by standard
Speech-and-Language Pathology (SLP) therapy.
Following SLP therapy, eligible patients still demonstrate aspiration
sufficient to require implantation of a tracheal tube. They must have sufficient cognitive and
physical ability to operate the stimulation unit. They must understand and consent to
participate in the experimental protocol.
This study is conducted
under an Investigational Device Exemption and has been approved by the
Institutional Review Board at
System
The experimental system is
composed of a Huntington Medical Research Institute bipolar helical electrode
on the recurrent laryngeal nerve (RLN) and a modified NeuroControl™
VOCARE™ stimulation system (Figure 1). The clinical safety of the electrode has been
previously demonstrated in devices for stimulation of the Vagus
nerve to control epilepsy [5, 7].
The NeuroControl
stimulation system has been safely implanted in more than 1,600 patients
worldwide. The modified implanted
receiver-stimulator (IRS) produced stimulation currents between 0.5 mA and
4 mA and pulse widths between 10 ms and 800 ms. All
power and stimulation information is transmitted via radio frequency link from
an external digital controller. The
controller generates a continuous stimulation train at 40 Hz, which has
been shown previously to produce vocal fold adduction (VFAd)
[8].
Procedure
Under general anesthesia, a tracheostomy was performed and the RLN was exposed via a
horizontal incision, two fingerbreadths above the sternal
manubrium. The
appropriate diameter helix electrode, either 1.5, 2.0 or 3.0 mm, was
selected to fit the nerve without constriction.
The electrode was carefully placed on the nerve and a loop of the
electrode lead was secured to the surrounding fascia to provide strain
relief. The IRS was implanted in the
chest on the same side as the electrode.
The leads were passed subcutaneously and connected to the IRS. After the radio frequency link to the IRS was
tested and VFAd was verified by observation of the
vocal folds, the surgical sites were closed.
Test Protocol
Following the surgery, the
patient remained in the hospital for three to seven days. While in the hospital, the patient’s vocal
folds were observed daily with the flexible fiber-optic laryngoscope. VFAd was verified
during stimulation by the implanted system.
Following the patient’s discharge from the hospital, the vocal folds
were observed twice per week via laryngoscope.
VFAd was observed with and without stimulation
and the parameters required to produce frank VFAd
were recorded.
The study protocol lasted
three months. During the first month,
the patient had a modified barium swallow (MBS) exam once each week. The MBS exam is the standard procedure for diagnosing
both patent and silent aspiration. It
provides a visual record of the food bolus in the different phases of a
swallow. During the second month, the
MBS was performed once every other week and at the beginning and the end of the
third month.
The patient’s swallow was observed
for three food consistencies: puree, thickened liquids, and thin liquids. The stimulation was OFF for half of the
presentations and ON for the other half.
Each swallow was recorded on videotape and later digitized. The recorded swallows were reviewed in random
order by a qualified SLP whom was blinded to whether stimulation was ON or
OFF. The SLP assessed each swallow for
aspiration in different phases of the swallow.
A statistical test of proportions tested the hypothesis that aspiration
was reduced with stimulation ON compared to stimulation OFF.
Results
To date, a single patient
has been enrolled in the study. The
patient is a 53 year-old female, eight years post-stroke.
During larygoscopic
examination, VFAd was verified in all trials. Frank VFAd was
defined as vocal cord closure that remained tightly closed during respiratory
cycles. The parameters for frank vocal
fold closure were approximately 1.0 mA with a pulse duration that varied
from a low of 56 ms and a high of 132 ms.
Typical stimulation pulse duration was 128 ms.
Anecdotally, it was noted that stronger stimulation would produce a strong cough. The patient could cough voluntarily, but only produced a single weak cough. If the stimulation pulse duration was turned up to 200 ms, the patient produced a series of very powerful coughs. The patient did not report any sensation, pain, or irritation during the stimulation.
The results of the MBS
trials are summarized in Table 1.
The gray sections of the table indicate significant differences (a < 0.1) between
stimulation ON and OFF. For all three
consistencies, the patient did not aspirate substantially during the
swallow. Rather the aspiration either occurred
prior to the swallow from premature entry into the laryngeal vestibule or
following the swallow from laryngeal or vallecular residue.
For puree and thickened liquids, the patient had very little aspiration
prior to and during the swallow. Without
stimulation, there was substantial aspiration after the swallow, which was significantly
reduced with stimulation (p < 0.1).
For the thin liquid, nearly
every trial without stimulation was aspirated prior to the swallow. This was a result of the patient’s poor oral
control leading to the bolus penetration past the vocal cords prior to
initiation of swallow. With stimulation
ON, the aspiration prior to swallow was significantly reduced (p < 0.1).
The last row in Table 1
shows the percentage of trials that demonstrated aspiration during any phase of
the swallow. Stimulation does not
significantly change aspiration for puree.
Stimulation does significantly reduced aspiration for both the thickened
liquid and thin liquid trials (p < 0.05).
Discussion
Stimulation of the RLN
always produced frank closure of the vocal folds, as observed through the
laryngoscope. By adjusting the
parameters, the vocal folds could be held closed during the inspiratory
and expiratory phases of breathing. This
supports the hypothesis that stimulation of the RLN produces VFAd. The parameters
to produce VFAd were stable throughout the three
months of the study.
The patient had poor oral
control of the food bolus. A portion of
the bolus would enter the laryngeal vestibule prior to the swallowing
effort. Thus it was important to maintain
VFAd for an extended period of time that usually
spanned several normal breaths.
The MBS results indicate that the patient’s aspiration occurred predominately before or after the swallowing effort, rather than during it. To prevent aspiration, stimulation must be applied before the bolus is placed in the mouth and remain ON until all food is cleared from the larynx and vallecular spaces. Timing the stimulation is straightforward during the MBS test as the food bolus is directly observed on the fluoroscope. In clinical or therapeutic use, it is difficult to determine the length and timing of stimulation because the bolus cannot be visualized. This could present an obstacle to clinical implementation of this therapy.
The strong cough produced by
increased stimulation was unexpected.
The ability to produce a strong cough could be beneficial as the
stimulated cough was significantly stronger than the patient’s voluntary
cough. The mechanism of cough generation
is unclear. One hypothesis is that
increased stimulation excites sensory fibers within the RLN and initiates a
reflexive response. A second hypothesis
is that the strong vocal fold closure injected a small amount of secretions
into the upper airway, resulting in the cough.
Results of the MBS exams show significant improvements overall for thickened and thin liquid, supporting the hypothesis that VFAd reduces aspiration. Liquids are typically the most difficult to control. Without stimulation, the patient aspirated 75% of the time on both liquids compared to only 44% on the puree. With stimulation the aspiration was reduced to 33%, 31%, and 50% with puree, thickened and thin liquid, respectively. While the stimulation did not eliminate aspiration, it did demonstrate a reduction. A reduction, however, may be sufficient as even normal individuals aspirate a small percentage of the time.
The least expected result was that stimulation did not change the percentage of aspiration during the swallow event. This indicates that aspiration resulted from poor oral control and remaining residue following the swallow. Stimulation of the RLN prior to and following the swallow protected the airway for an extended period of time compared to no stimulation, but did not change the airway protection during swallow. Qualitative analysis of the MBS results demonstrates better bolus clearance and less residual with stimulation compared to without stimulation.
This is only the first of several patients. While a single patient is insufficient to prove our hypotheses, the results from this initial patient demonstrate the feasibility and promise of RLN stimulation to control aspiration.
References
[1] R. W. Teasell, M. McRae, Y. Marchik, and H. M. Finestone,
"Pneumonia Associated with Aspiration Following Stroke," Archives of Physical Medicine and
Rehabilitation, vol. 77, pp. 707-709, 1996.
[2] J. E. Aviv, J. H. Martin, R. L. Sacco, D. Zagar, B. Diamond, M.
S. Keen, and A. Blitzer, "Supraglottic and
pharyngeal sensory abnormalities in stroke patients with dysphagia,"
Ann Otol Rhinol Laryngol, vol. 105,
pp. 92-7, 1996.
[3] M. Broniatowski, R. Dessoffy, K. Azar, C. R. Davies,
M. S. Trott, F. R. Miller, and H. M. Tucker,
"Electronic integration of glottic closure and cricopharyngeal relaxation for the control of aspiration: A
canine study," Otolaryngology-Nead and Neck Surgery, vol. 112, pp. 424-429, 1995.
[4] M. Broniatowski, L.
A. Ilyes, G. B. Jacobs, Y. Nose, and H. M. Tucker,
"Artificial reflex arc: A potential solution for chronic aspiration
II. A canine study based on laryngeal
prosthesis," Laryngoscope, vol.
98, pp. 235-237, 1988.
[5] W. B. Tarver, R. E. George, S. E. Maschino, L. K. Holder, and J. F. Wernicke,
"Clinical experience with a helical bipolar stimulating lead," Pacing Clin Electrophysiol, vol. 15, pp. 1545-56, 1992.
[6] A. Agur, Grant's Atlas of Anatomy, 9th Edition
ed. Baltimore, MD: Williams & Wilkins.
[7] R. S. Terry, W. B. Tarver, and J. Zabara, "The implantable neurocybernetic
prosthesis system," Pacing Clin Electrophysiol, vol. 14,
pp. 86-93, 1991.
[8] D. S. Lundy, R. R. Casiano,
H. J. Landy, J. Gallo, B. Gallo, and R. E. Ramsey,
"Effects of vagal nerve stimulation on laryngeal
function," J Voice, vol. 7, pp.
359-64, 1993.
Acknowledgments
The work was supported
in part by NIH SBIR grant number 1 R43 NS38776‑01,
NeuroControl Corp., and St. Vincent’s Charity
Hospital (University Hospital Health System).