Efficacy of Combined Inspiratory Intercostal and Expiratory Muscle Pacing to Maintain Artificial Ventilation

Efficacy of Combined Inspiratory Intercostal and Expiratory Muscle Pacing to Maintain Artificial Ventilation ANTHONY F. DIMARCO, JAROSLAW R. ROMANIUK, KRZYSZTOF E. KOWALSKI, and GERALD S. SUPINSKI Pulmonary Division, Department of Medicine, Case Western Reserve University, MetroHealth Medical Center, Cleveland, Ohio Many patients with ventilator-dependent quadriplegia have coincident phrenic nerve injury and therefore cannot be offered phrenic nerve pacing. The purpose of this study was to assess the utility of combined inspiratory intercostal and expiratory muscle pacing to provide complete ventilatory support. Studies were performed in 15 anesthetized dogs. An electrode was positioned on the epidural surface of the upper thoracic spinal cord to activate the inspiratory intercostal muscles; a separate electrode was positioned on the epidural surface of the lower thoracic spinal cord to activate the expiratory muscles. In an attempt to replicate the effects of inspiratory intercostal pacing alone in humans, stimulus parameters during upper thoracic spinal cord stimulation were adjusted to provide suboptimal levels of ventilation (end-tidal PCO 2 of 55 to 60 mm Hg). Expiratory muscle activation was triggered electrically by the inspiratory signal with a 4.2-s delay resulting in alternate inspiratory and expiratory muscle pacing at a combined rate of 14 breaths/min. Combined pacing was maintained for an arbitrary period of 3 h. Initial intercostal muscle pacing alone resulted in an end-tidal PCO 2 of 57.1 1.1 mm Hg. After the addition of expiratory muscle pacing, end-tidal PCO 2 fell to 36.3 1.2 mm Hg. Tidal volume during both inspiratory and expiratory muscle pacing and end-tidal PCO 2 remained stable throughout the study period. Our results suggest that combined alternate inspiratory and expiratory muscle pacing may be a viable alternative method of artificial ventilation in ventilatordependent quadriplegic patients. DiMarco AF, Romaniuk JR, Kowalski KE, Supinski GS. Efficacy of combined inspiratory intercostal and expiratory muscle pacing to maintain artificial ventilation. AM J RESPIR CRIT CARE MED 1997;156:122 126. Phrenic nerve pacing allows patients with ventilator-dependent quadriplegia freedom from mechanical ventilation and improvement in their overall quality of life (1 5). Unfortunately, many patients with cervical spinal cord injury have insufficient phrenic nerve function because of coincident injury to their phrenic motoneuron pools and/or phrenic nerves and consequently cannot be offered this technique. In previous human studies, we have demonstrated that the intercostal muscles can be activated via upper thoracic spinal cord stimulation (SCS) and that this technique results in substantial inspired volume production (6). This technique alone, however, does not result in sufficient inspired volumes to support ventilation for prolonged time periods (6). We hypothesized that partial ventilatory support could also be provided by contraction of the expiratory muscles and that this method could be used to augment ventilation provided by (Received in original form September 19, 1996 and in revised form February 3, 1997 ) Supported by Grant No. HL-34143 from the National Institutes of Health. Correspondence and requests for reprints should be addressed to A. F. DiMarco, M.D., MetroHealth Medical Center, 2500 MetroHealth Drive, Suite 323, Cleveland, OH 44109. Am J Respir Crit Care Med Vol. 156. pp. 122 126, 1997 inspiratory intercostal pacing. Intermittent expiratory muscle contraction would result in exhalation below FRC followed by passive inhalation. In this regard, we have previously demonstrated that a major portion of the expiratory muscles can be reproducibly activated via lower thoracic SCS (7). The purpose of this study, therefore, was to evaluate the utility of combined inspiratory intercostal and expiratory muscle pacing to support artificial ventilation in anesthetized dogs. To simulate the effects of intercostal pacing alone in humans (6), upper thoracic SCS was initially adjusted to produce suboptimal levels of ventilation, resulting in significant hypercapnia. Expiratory muscle pacing was then added to inspiratory intercostal pacing in attempt to restore eucapnia. METHODS Studies were performed in 15 mongrel dogs (weight, 14.5 to 20.9 kg; mean, 17.3 0.5 kg) lying in the supine posture. All animals were anesthetized with pentobarbital (initial dose, 35 mg/kg), and then additional doses of PB (3 to 5 mg/kg/h) as required to suppress any response to painful stimuli, the corneal reflex, and spontaneous respirations. Each animal was intubated with a large-bore (10 mm ID), cuffed endotracheal tube. A homeothermic blanket (Harvard Apparatus, Cambridge, MA) was used to maintain body temperature at 38 0.5 C. End-tidal PCO 2 was monitored with a rapidly responding CO 2 analyzer (O.R. SARAcap; PPG Biomedical Systems, Lenexa, KS) at the tracheal opening. The flow signal from a pneumotachograph

DiMarco, Romaniuk, Kowalski, et al.: Muscle Pacing in Artificial Ventilation 123 (Fleisch No. 1; OEM Medical, Richmond, VA) was integrated to obtain tidal volume. Tracheal pressure was measured with a differential pressure transducer (Validyne MP-45; Validyne, Northridge, CA). A cannula was placed in the femoral artery to allow blood pressure monitoring (P23XL, SpectraMed; Statham Instruments, Hato Rey, PR) and blood withdrawal for measurement of arterial blood gases (Arterial Blood Gas Analyzer, ABL-30; Radiometer, Copenhagen, Denmark). A separate cannula was placed in the femoral vein for the administration of supplemental anesthesia. The cervical phrenic rootlets (C 5 C 7 ) were sectioned bilaterally. An EM Recorder (Electronics for Medicine, White Plains, NY) was used to record all signals. A flexible platinumiridium unipolar electrode (Medtronic, Inc., Minneapolis, MN) was inserted onto the epidural surface of the spinal cord via a midthoracic laminectomy and advanced to the T 2 spinal cord level to activate the inspiratory intercostal muscles. The precise location that resulted in maximal negative pressure generation during electrical stimulation against an occluded airway was used to determine final electrode position, according to previously described techniques (8, 9). A separate stimulating electrode was positioned epidurally at the T 9 T 10 spinal cord region to activate the expiratory muscles (7). Likewise, the precise location that resulted in maximum positive pressure generation during stimulation against an occluded airway was used to determine the final position of this electrode. A two-channel electrical stimulator (Model S88; Grass Instruments, Quincy, MA) was used to activate the inspiratory intercostal and expiratory muscles, via upper and lower thoracic SCS, respectively. After hyperventilation-induced apnea, supramaximal stimulus parameters were determined in each animal, i.e., the stimulus amplitudes and frequencies that resulted in maximal negative and positive swings in airway pressure (Pmax) during upper and lower thoracic SCS, respectively. Stimulation was always applied under conditions of airway occlusion. Pressure-frequency curves for the expiratory muscles were subsequently obtained by determining the change in airway pressure during stimulation, utilizing supramaximal stimulus amplitudes at several different stimulus frequencies over the 10 to 50 Hz range. Finally, the change in airway pressure during airway occlusion, using the submaximal stimulus parameters selected for prolonged inspiratory intercostal and expiratory muscle pacing ( P) were also determined in each animal. These values were used to calculate the fraction of maximal pressure ( P/Pmax) required during both inspiratory intercostal and expiratory muscle pacing. In each animal, the product of P/Pmax and duty cycle (inspiratory or expiratory time/total cycle duration [Ttot]) or pressure-time index was calculated. Protocol Because intercostal pacing alone in humans does not result in sufficient levels of ventilation to maintain eucapnia, we sought to simulate this condition during intercostal pacing alone in our animal model. In addition, we sought to achieve a relatively normal final respiratory rate during combined alternate inspiratory and expiratory muscle pacing. Intercostal pacing, therefore, was initially set at a rate of 7 breaths/ min. Stimulus frequency was arbitrarily set at 20 Hz, and adjustments in stimulus amplitude were then made to achieve a tidal volume that resulted in an end-tidal PCO 2 of 55 to 60 mm Hg. Inspiratory intercostal muscle pacing was maintained at this level for approximately 30 min. Subsequently, expiratory muscle pacing was added at the same respiratory rate of 7 breaths/min. Expiratory muscle activation was triggered electrically by the inspiratory signal with a 4.2-s delay. Stimulus frequency during expiratory muscle activation was also set at 20 Hz; stimulus amplitude was adjusted to the lowest stimulus amplitude necessary to maintain end-tidal PCO 2 at approximately 40 mm Hg. Combined alternate inspiratory and expiratory muscle pacing was then maintained continuously for an arbitrary period of 3 h. End-tidal PCO 2 was continuously monitored in each animal; arterial blood gases were monitored intermittently during the 3.5-h study period in nine animals. Inspired volumes and minute ventilation were monitored every 30 to 60 min. Pressure-frequency responses of the expiratory muscles were also monitored every hour. Bipolar recording electrodes were placed in the medial portion of the parasternal intercostal muscle (third intercostal space) in each animal; electrical pacing was intermittently discontinued to verify the absence of spontaneous respiratory activity. Additional pentobarbital anesthesia (2 to 3 mg/kg intravenously) was administered, if necessary. Statistical analysis was performed using a one-way analysis of variance and the Newman-Keuls test, when applicable. A p value of 0.05 was taken as significant. RESULTS An example of inspiratory intercostal muscle pacing alone and combined alternate inspiratory intercostal and expiratory muscle pacing is shown in Figure 1. Inspiratory intercostal pacing is characterized by the monotonous generation of inspired volumes of 280 ml, at an interval of 8.4 s (7 breaths/ min). In this example, this breathing pattern resulted in an end-tidal PCO 2 of 58 mm Hg. With combined alternate pacing Figure 1. Breathing pattern during inspiratory muscle pacing alone (upper panel) and combined inspiratory and expiratory muscle pacing (lower panel). Stimulus artifact was recorded from the parasternal intercostal electrode. See text for further description.

124 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL. 156 1997 Figure 2. Mean tidal volumes generated during chronic respiratory muscle pacing. There were no significant changes in volume production for either inspiratory or expiratory muscle pacing over the 3.5-h period of study. (Figure 1, lower panel), the expiratory muscles were activated after each breath generated by inspiratory muscle contraction such that a regular respiratory pattern was established at a breathing frequency of 14 breaths/min. The addition of expiratory muscle pacing resulted in a fall in end-tidal PCO 2 to 38 mm Hg. The mean tidal volumes generated during respiratory muscle pacing over the 3.5-h time period of the study are shown in Figure 2. Initial inspired volumes were 279 13 and 218 9 ml for separate inspiratory and expiratory muscle activation, respectively, and they remained stable during the period of chronic pacing. For example, during the final 30 min of pacing, inspired volumes were 263 12 and 207 9 ml for separate inspiratory and expiratory muscle activation, respectively (NS). The effects of respiratory muscle pacing on end-tidal PCO 2 were monitored in each animal throughout the period of stimulation. Mean end-tidal PCO 2 during inspiratory intercostal muscle pacing alone was 57.1 1.1 mm Hg; after the addition of expiratory muscle pacing, mean end-tidal PCO 2 fell to 36.3 1.2 mm Hg (p 0.001). Combined alternate inspiratory and expiratory muscle pacing resulted in stable values of end-tidal PCO 2 throughout the 3-h study period. At the completion of the 3-h time period, PCO 2 was 36.4 1.6 mm Hg (NS). Arterial blood gas results are presented in Table 1. Initial mean Pa O 2 was 64.5 mm Hg during inspiratory intercostal pacing alone. After the addition of expiratory muscle pacing, Pa O 2 increased to 89.1 mm Hg (p 0.01), and it did not change significantly during the 3-h period of combined pacing. Arterial Pa CO 2 values were very similar to the end-tidal PCO 2 values. Spontaneous parasternal activity was absent during intercostal pacing alone. After intermittent discontinuation of pacing during combined pacing, animals remained apneic for at least 20 s as determined by the absence of parasternal EMG activity. Ventilatory parameters and mean airway pressure generation during airway occlusion during maximal stimulation (Pmax) and with chronic pacing parameters ( P) are provided in Table 2. Because of the much higher Pmax during lower thoracic SCS and relatively similar P during chronic stimulation, P/Pmax during lower thoracic SCS was significantly lower (0.29 0.03) than that for upper thoracic SCS (0.59 0.02). Moreover, because the duty cycles were identical during upper and lower thoracic SCS, the pressure-time index was significantly lower during lower thoracic SCS (0.03) compared with that during upper thoracic SCS (0.07 0.01). The mean relationship between stimulus frequency and airway pressure generation at the start of the first and during the final hour of electrical stimulation for lower thoracic SCS is shown in Figure 3. There was a small downward and rightward shift of this curve, which was not statistically significant, suggesting the absence of significant system fatigue during prolonged pacing. DISCUSSION In previous studies (8), we have demonstrated that inspiratory intercostal muscle pacing via upper thoracic SCS alone in a dog model is sufficient to maintain eucapnia for prolonged time periods. In a subsequent study, we also demonstrated TABLE 1 MEAN GAS TENSIONS DURING CHRONIC PACING* Time Inspiratory muscle pacing alone 0.5 h Pa CO 2 58.0 1.5 Pa O 2 64.5 8.9 0 h 55.7 1.4 63.3 8.5 Combined inspiratory and expiratory muscle pacing 1 h 36.6 1.5 89.7 9.1 2 h 36.3 1.6 83.8 7.3 3 h 36.1 1.9 83.4 6.7 * Values are mean SE; n 9. Significantly lower than inspiratory pacing alone (p 0.001). Significantly higher than inspiratory pacing alone (p 0.01).

DiMarco, Romaniuk, Kowalski, et al.: Muscle Pacing in Artificial Ventilation 125 TABLE 2 MEAN PRESSURE AND TIMING INDICES EMPLOYED DURING PROLONGED INSPIRATORY AND EXPIRATORY MUSCLE STIMULATION* Inspiratory (upper SCS) and Expiratory (lower SCS) Time (s) Total Cycle Time (s) Duty Cycle Airway Pressure during Tracheal Occlusion Chronic Stimulation (cm H 2 O) Airway Pressure during Tracheal Occlusion Supramaximal Stimulation (cm H 2 O) P/Pmax Pressure-Time Index Upper SCS 1.0 8.4 0.12 10.8 0.5 18.3 0.7 0.59 0.02 0.07 0.01 Lower SCS 1.0 8.4 0.12 14.2 0.9 52.3 3.9 0.29 0.03 0.03 0.01 Definition of abbreviations: SCS spinal cord stimulation; P/Pmax fraction of maximal pressure. * Values are mean SE; n 15. Duty cycle times pressure change/pmax. Significantly lower than upper thoracic SCS (p 0.001). Significantly higher than upper thoracic SCS (p 0.001). that this technique results in substantial inspired volume generation in ventilator-dependent quadriplegic patients as well (6). Unfortunately, this technique does not result in sufficient tidal volume generation to maintain adequate levels of ventilation in humans for more than a few hours (6). Because the enervation of the expiratory muscles is intact in most of these patients and these muscles can potentially be activated via lower thoracic SCS (7), our results suggest that combined alternate inspiratory and expiratory muscle pacing may be a viable alternative method of artificial ventilation in this patient population. During this type of study, spontaneous inspiratory or expiratory muscle contraction could make an unrecognized contribution to tidal volume production. It was important, therefore, to ensure that spontaneous respiratory activity was completely suppressed. Our data suggest that we were successful in this regard. First, all animals were significantly hypercapnic during inspiratory intercostal muscle pacing alone, reflective of the deep anesthetic level maintained throughout Figure 3. Relationship between stimulus frequency and airway pressure generation at the start of the first (closed circles) and during the final hour (open diamonds) of electrical stimulation for lower thoracic SCS. See text for further explanation. the study period. In addition, discontinuation of electrical stimulation resulted in periods of apnea, as demonstrated by absent parasternal intercostal EMG activity, for at least 20 s. In previous studies (7), we have shown that the application of electrical current in the T 9 T 10 region of the lower thoracic spinal cord results in the activation of a major portion of the expiratory muscles and the consequent generation of large positive airway pressures. Although the precise pathway of current spread is unknown, this technique results in the activation of the internal intercostal muscles of the lower rib cage as well as the abdominal muscles (10). There is no apparent activation of the limb musculature. The external intercostal muscles of the lower rib cage, conventionally considered inspiratory muscles, however, are also activated. These muscles are extremely thin (11, 12), however, and therefore did not likely have any adverse impact on exhaled volume or positive airway pressure generation during lower thoracic SCS. The abdominal muscles, the major group of expiratory muscles, are generally engaged in tasks requiring rapid, forceful contractions of short duration such as coughing or sneezing (13). In contrast, tidal volume generation, a repetitive task, requires high levels of endurance. However, the abdominal muscles such as the diaphragm have a relatively high proportion (approximately 50%) of slow-twitch, highly oxidative (Type I) fibers (14 16). The expiratory intercostal muscles have an even higher proportion of Type I fibers (approximately 60%) (17). Capillary supply to the expiratory intercostal muscles is also high, even greater than that for the diaphragm (14). On the basis of their morphology alone, therefore, it is conceivable that these muscles could adequately perform the role of inspiratory muscles. The prestimulation fiber-type characteristics and endurance properties of the respiratory muscles, however, may not be relevant. Previous studies have indicated that activation of skeletal muscle with repetitive low frequency electrical stimulation over prolonged time periods results in the remodeling of fast muscles to predominantly slow type I fibers, which are highly oxidative and fatigue-resistant (16, 18). Previous investigations have described a critical tensiontime index (product of percent maximal transdiaphragmatic pressure and duty cycle) for the human diaphragm below which fatigue does not occur (19). Breathing patterns characterized by tension-time indices of less than 0.15 could be maintained for at least 45 min, whereas values above this level resulted in the development of diaphragm fatigue. In previous work (8), we found that intercostal pacing alone with stimulus parameters sufficient to maintain artificial ventilation for several hours could be achieved with a calculated pressure-time

126 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL. 156 1997 index ( P/Pmax Ti/Ttot) of 0.12. No significant evidence of system fatigue was evident over this time period. It was not surprising, therefore, that intercostal pacing in the present study, utilizing approximately one-half the duty cycle of the previous study and similar P/Pmax, did not result in any appreciable development of system fatigue. The calculated pressure-time index for the expiratory muscles in the present study was even lower than that for the inspiratory intercostal muscles (approximately 0.03), in large measure secondary to the much greater Pmax of the expiratory muscles. It should be mentioned, however, that the critical pressure-time index of the expiratory muscles is unknown and may be quite different from that for the diaphragm. Nevertheless, at the breathing patterns used in the present study, there was no significant evidence of fatigue during the 3-h time period of the study, suggesting that the magnitude of force development during pacing and pacing frequency were performed below any fatiguing threshold for these muscles. Phrenic nerve pacing in patients with ventilator-dependent quadriplegia results in improved speech, increased patient comfort, greater ease with patient transport, and overall improved quality of life (1 5). In patients with only partial phrenic nerve function, and therefore not candidates for phrenic nerve pacing, we have shown that combined inspiratory intercostal and diaphragm pacing can also provide sufficient ventilation to maintain full-time ventilatory support (20). Moreover, this technique appears to offer these patients similar benefits to those of a bilateral diaphragm pacing. The results of the present study demonstrate that expiratory muscle pacing can also provide significant ventilatory support and, when combined with other techniques, could be a useful modality in several different subgroups of patients with ventilator-dependent quadriplegia. As used in the present study, expiratory muscle pacing could be combined with inspiratory intercostal muscle pacing in patients with absent phrenic nerve function. Alternatively, expiratory muscle pacing could be combined with phrenic nerve pacing in those patients with only partial phrenic nerve function and therefore not candidates for phrenic nerve pacing alone. Finally, this technique may be useful in spontaneously breathing patients with spinal cord injuries and suboptimal diaphragm function and consequent marginal tidal volumes, particularly while in the sitting posture. References 1. Glenn, W. W. L., J. F. Hogan, J. S. O. Loke, T. E. Ciesielski, M. L. Pheps, and R. Rowedder. 1984. Ventilatory support by pacing of the conditioned diaphragm in quadriplegia. N. Eng. J. Med. 310:1150 1155. 2. Glenn, W. W. L. 1980. The treatment of respiratory paralysis by diaphragm pacing. Ann. Thorac. Surg. 30:106 109. 3. Shneerson, J. M., and J. Moxham. 1993. Diaphragm pacing. Am. Rev. Respir. Dis. 148:533 536. 4. Glenn, W. W. L., and H. Sairenji. 1985. Diaphragm pacing in the treatment of chronic ventilatory insufficiency. In C. Roussos and P. T. Macklem, editors. The Thorax, Part B. Marcel Dekker, New York. 1407 1440. 5. Miller, J. I., J. A. Farmer, W. Stuart, and D. Apple. 1990. Phrenic nerve pacing of the quadriplegic patient. J. Thorac. Cardiovasc. Surg. 99:35 40. 6. DiMarco, A. F., G. S. Supinski, J. A. Petro, and Y. Takaoca. 1994. Evaluation of intercostal pacing to provide artificial ventilation in quadriplegics. Am. J. Respir. Crit. Care Med. 150:934 940. 7. DiMarco, A. F., J. R. Romaniuk, and G. S. Supinski. 1995. Electrical activation of the expiratory muscles to restore cough. Am. J. Respir. Crit. Care Med. 151:1466 1471. 8. DiMarco, A. F., K. Budzinska, and G. S. Supinski. 1989. Artificial ventilation by means of electrical activation of the intercostal /accessory muscles alone in anesthetized dogs. Am. Rev. Respir. Dis. 139:961 967. 9. DiMarco, A. F., M. D. Altose, and A. Cropp. 1987. Activation of the inspiratory intercostal muscles by electrical stimulation of the spinal cord. Am. Rev. Respir. Dis. 136:1385 1390. 10. DiMarco, A. F., J. R. Romaniuk, K. E. Kowalski, and G. S. Supinski. 1996. Pattern of expiratory muscle activation via spinal cord stimulation (abstract). Am. J. Respir. Crit. Care Med. 153:A21. 11. Miller, M. E., G. C. Christensen, and E. Evans. 1964. Anatomy of the Dog. W. B. Saunders, Philadelphia. 12. Woodburne, R. T. 1969. Essentials of Human Anatomy. Oxford University Press, New York. 13. Campbell, E. J. M., E. Agostoni, and J. Newsom-Davis. 1970. The Respiratory Muscles: Mechanics and Neural Control. Lloyd-Luke, London. 14. Mizuno, M. 1991. Human respiratory muscles: fiber morphology and capillary supply. Eur. Respir. J. 4:587 601. 15. Häggemark, T., and A. Thorstensson. 1979. Fiber types in human abdominal muscle. Acta Physiol. Scand. 107:319 325. 16. Phillips, C. A., J. S. Petrofsky, D. M. Hendershot, and D. Stafford. 1984. Functional electrical exercise: a comprehensive approach for physical conditioning of spinal cord injured patients. Orthopedics 7:1112. 17. Reid, M. B., G. C. Ericson, K. A. Feldman, and R. L. Johnson. 1987. Fiber types and fiber diameters in canine respiratory muscles. J. Appl. Physiol. 62:1705 1712. 18. Ciesielski, T. E., Y. Fukuda, W. W. L. Glenn, J. Gortien, K. Jeffrey, and J. F. Hogan. 1983. Response of the diaphragm muscle to electrical stimulation of the phrenic nerve: a histochemical and ultrastructural study. J. Neurosurg. 58:92 100. 19. Bellemare, F., and A. Grassino. 1982. Effect of pressure and timing of contraction on human diaphragm fatigue. J. Appl. Physiol. 53:1190 1195. 20. DiMarco, A. F., J. R. Romaniuk, K. Kowalski, and G. S. Supinski. 1995. Efficacy of combined inspiratory intercostal and expiratory muscle pacing to maintain artificial ventilation (abstract). Am. J. Respir. Crit. Care Med. 151:A417.