Hemidiaphragmatic Paresis During Interscalene Brachial Plexus Block: Effects on Pulmonary Function and Chest Wall Mechanics William F. Urmey,
MD,
and Marianne McDonald,
BS, RRT
Department of Anesthesiology, The Hospital for Special Surgery, Cornell University Medical College, New York, New York
We studied the effects of unilateral hemidiaphragmatic paresis caused by interscalene brachial plexus block on routine pulmonary function in eight patients. In an additional four patients, we studied changes in chest wall motion during interscalene block anesthesia by chest wall magnetometry. Ipsilatera1 hemidiaphragmatic paresis, as diagnosed by ultrasonography, developed in all patients within 5 min of interscalene injection of 45 mL of 1.5% mepivacaine with added epinephrine and bicarbonate. Large decreases in all pulmonary function variables were measured in every patient. Forced vital capacity and forced expiratory volume at 1 s de-
B
rachial plexus anesthesia by interscalene injection of local anesthetic routinely causes sensory anesthesia of the fourth and fifth, and frequently of the third, cervical nerve. Using radiographic contrast labeling, Winnie (1) demonstrated that this was due to local anesthetic entering the cervical plexus after interscalene brachial plexus injections as small as 10 mL. Local anesthetic entering the cervical plexus at these levels probably results in one-sided diaphragmatic paresis, as the C-3-C-5 nerve roots form the phrenic nerve. In a prior investigation of 13 healthy patients (2), we found that interscalene block resulted in a 100% incidence of ipsilateral hemidiaphragmatic paresis. Diagnosis was made by ultrasonography (3-5), which clearly demonstrated paradoxical motion of the cephalad border of the zone of apposition of diaphragm to rib cage on the ipsilateral side during forceful sniff (a maneuver that rapidly decreases pleural pressure). Previous studies have shown consistent sizeable reductions in pulmonary f unction during pathological Accepted for publication October 21, 1991. Address correspondence to Dr. Urmey, Department of Anesthesiology, The Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021.
352
Anesth Analg 1992;74:352-7
creased 27% 2 4.3% and 26.4% f 6.8%,respectively (P = 0.0001). Peak expiratory and maximum midexpiratory flow rates were also significantly reduced. Interscalene block caused changes in pulmonary function and chest wall mechanical motion that were similar to those published in previous studies on patients with hemidiaphragmatic paresis of pathological or surgical etiology. Interscalene block probably should not be performed in patients who are dependent on intact diaphragmatic function and in those patients unable to tolerate a 25%reduction in pulmonary function. (Anesth Analg 1992;74352-7)
unilateral phrenic nerve paralysis (6,7) as well as during direct local anesthetic blockade of one phrenic nerve (8). We determined that alterations in routine pulmonary function tests and chest wall mechanics occur as the result of unilateral hemidiaphragmatic paresis produced by interscalene block and how these pulmonary function test changes might compare to those previously published during complete unilateral pathological phrenic nerve paralysis. Chest wall magnetometry was used to characterize the physiologic alterations in chest wall motion that occurred during the hemidiaphragmatic paresis that resulted from interscalene block.
Methods After approval from our Institutional Review Board and written informed consent had been obtained, eight consecutive adult patients were studied. Patients were ASA class 1-111 and had been scheduled for upper extremity orthopedic procedures under interscalene brachial plexus anesthesia. Two patients had preexisting pulmonary disease (patient 3, moderate chronic obstructive pulmonary disease and patient 7, asthma). No premedication or sedation was 61992 by the International Anesthesia Research Society 0003-2999/92/$5.00
ANESTH ANALG 1992;74:352-7
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATICPARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
given until all testing was completed. Interscalene block was administered with the needle tip at, or caudad to, cricoid level according to the method of Winnie (1). A single paresthesia technique with a 23-gauge 2.5-cm needle was used. Each patient received 45 mL of 1.5% mepivacaine with 5 pg/mL of epinephrine and 0.05 mEq/mL of sodium bicarbonate. Before anesthesia, real-time ultrasonography (Diasonics DRF-400, 3 - m transducer) ~ of the ipsilateral hemidiaphragm was done to c o n f m normal caudad motion during sniff. After application of noseclips, data were collected during two successive forced vital capacity maneuvers into a computerized bell spirometer (Warren E. Collins, Eagle II, Braintree, Mass.) with coaching by a respiratory therapist. Expiratory and inspiratory flow-volume curves were then generated by forced expiration and inspiration into a Wedge Spirometer (Med Science, St. Louis, Mo.). X-Y flow-volume data were recorded from a dual-trace storage oscilloscope (Tektronix, Beaverton, Ore.) for later analysis. These measurements were repeated 5 min after completion of interscalene local anesthetic injection, by which time all patients had ultrasonographically diagnosed ipsilateral hemidiaphragmatic paresis (paradoxical cephalad hemidiaphragmatic motion during s n q . To determine whether a reduction in expiratory function occurred during hemidiaphragmatic paresis that was not based on a reduction in total lung capacity, data were analyzed by volume-indexing the flow-volume curves. This was done by superimposing the curves at the residual volume (RV) points and comparing the postanesthesia peak expiratory flow rate (PEFR) with the baseline flow rate at the same lung volume, as shown in Figure 1. To analyze the changes in chest wall mechanical motion that occur during interscalene block, an additional four patients were studied using chest wall magnetometry (GMG Scientific, Boston, Mass.). Measurements of anteroposterior diameter changes of the rib cage at midsternum, of the left of the abdomen, and separately of the right half of the abdomen at the umbilical level were made during tidal and deep breathing (Figure2). Data were recorded immediately before and 15min after interscalene block and 10 min after ultrasonographically diagnosed hemidiaphragmatic paresis. X-Y plots of these data were photographed from the dual-trace storage oscilloscope.
Results Ultrasonographic examination revealed that in all eight patients normal ipsilateral hemidiaphragmatic caudad motion changed to paradoxical cephalad motion during forceful sniff within 2 min of interscalene
353
Figure 1. Representative data from a healthy patient comparing baseline (before interscalene block) maximum expiratory flowvolume curve with those measured at 2 and 5 min after interscalene injection of 45 mL of 1.5%mepivacaine with epinephrine. Decreased peak flows at same volumes (referenced to residual volume, RV) constitute evidence that the intact diaphragm normally contributes to forced expiration. TLC represents total lung capacity before interscalene block. See text for discussion.
Rib
Cagr AP Memrter
Figure 2. Representative modified Konnc-Mead diagram is shown for one of the study patients. Vertical axis, ipsilateral and contralateral hemiabdominal anteroposterior (AP) diameter; horizontal axis, rib cage anteroposterior diameter. Tidal breathing loops beginning at functional residual capaaty (FRC) are shown. Ipsilateral hemiabdominal expansion during inspiration changed to contraction (paradoxical motion) during interscalene block in every patient. Contralateral hemiabdominal expansion during inspiration was more preserved. All patients had comparatively more rib cage than abdominal expansion with inspiration during interscalene anesthesia, reflecting recruitment of intercostal and accessory muscles of inspiration to compensate for compromised diaphragmatic inspiratory action. Solid line, before anesthesia; dashed line, during anesthesia.
anesthetic injection. Thus, ipsilateral hemidiaphragmatic paresis developed in 100% of the patients studied, as had been observed in our previous study. Spirometry results for patients 3 and 7 who had preexisting obstructive lung disease showed changes during anesthesia that were similar in value to those measured in the six healthy patients. Table 1 displays the results of routine pulmonary function testing. All eight patients had large reduc-
354
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
ANESTH ANALG 1992;74:352-7
Table 1. Comparison of Pulmonary Function Measurements Before and During Interscalene Block Spirometry Patient No. 1
2 3 4 5 6
7 8
Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change
Flowlvolume loop
FVC (L)
FEV1.0 (L)
PEFR (Us)
FEF,,, (Us)
PEFR (Lls)
PIFR (Us)
3.36 2.50 -25.60 3.73 2.45 -34.32 5.63 4.00 -28.95 5.96 4.19 -29.70 4.14 2.94 -28.99 4.52 3.58 -20.80 2.41
3.02 2.20 -27.15 3.32 2.17 -34.64 4.43 3.32 -25.06 4.11 3.04 -26.03 3.09 2.36 -23.62 3.57 2.97 -16.81 1.27 0.80 -37.01 3.98 3.17 -20.35
7.39 6.06 -18.00 6.50 5.25 -19.23 11.80 10.40 -11.86 10.00 7.60 -24.00 5.18 5.12 -1.16 6.99 7.72 10.44 2.37 1.36 -42.62 8.70 7.32 -15.86
3.73 2.52 -32.44 4.22 2.88 -31.75 3.85 3.09 -19.74 2.62 2.03 -22.52 2.48 2.47 -0.40 3.37 3.07 -8.90 0.47 0.41 -12.77 4.14 3.56 -14.01
6.0 5.1 -15.0 5.6 5.5 -1.8 8.0 7.9 -1.3 7.4 6.5 -12.2 5.8 4.5 -22.4 8.1 8.4 3.7 1.3 1.3 0.0 7.0 7.0 0.0
3.8 3.8 0.0 5.9 3.3 -44.9 7.5 7.3 -2.7 4.0 3.9 -2.5 5.0 3.8 -24.0 5.5 5.8 5.5 1.0 1.8 80.0 7.0 6.0 -14.3
1.77 -26.56 4.77 3.70 -22.43
FVC, forced vital capacity; FEV,,,, forced expiratory volume at 1 s; PEFR, peak expiratory flow rate; FEF2s75, forced expiratory flow hom 25% to 75%of the vital capacity; PIFR, peak inspiratory flow rate.
tions in forced vital capacity (FVC) ranging from 21% to 34% within 5 min of interscalene local anesthetic injection. Mean FVC decreased 27.2% t 4.3% (P = 0.0001) for the group. Similarly, forced expiratory volume at 1 s (FEVl.o)decreased between 17% and 37%, with a mean reduction of 26.4% 2 6.8% (P = 0,0001). Postanesthesia peak expiratory flow rate decreased in all but one patient, with a mean reduction of 15.4% t 15.7% (P < 0.05). Maximum midexpiratory flow rate from 25% to 75% of the vital capacity decreased in every patient with a mean diminution of 17.9% t 11.1% (P = 0.05). Volume-indexed PEFR decreased 6.0% t 9.1% (P = 0.05) during interscalene block (Figure 1). Inspiratory flow-volume curve analysis showed that peak inspiratory flow rates declined in five of eight patients and was unchanged in one patient. However, these data were less reproducible (Table l)than expiratory flow and volume data. Using magnetometry, ipsilateral and contralateral hemiabdominal anteroposterior diameters were plotted as separate functions of increasing rib cage anteroposterior diameter during spontaneous inspiration in four patients. A representative tracing from one of the patients is shown in Figure 2. After interscalene block, rib cage motion was more pronounced than abdominal motion in all four patients.
Because of the nature of the measurements, anteroposterior diameter changes by magnetometry may be small during normal or deep breathing as one chest wall compartment may not move at all during breathing. However, the direction of motion and comparison of the chest wall compartments may yield valuable information regarding chest wall muscle function during breathing. In this case, the increase in rib cage activity reflected heightened activity of intercostal and accessory muscles of inspiration (muscles that act on the rib cage and not the abdomen), which are recruited to take over for the compromised diaphragm (9). In addition, paradoxical (inward) ipsilateral hemiabdominal motion during quiet or deep inspiration developed in each patient after receiving interscalene block. By contrast to the ipsilateral side, contralateral hemiabdominal expansion was diminished in three of four patients (and actually increased in the one illustrated patient), but normal directional motion was preserved in all four patients. Thus, the abdomen departed from its usual one degree of freedom of motion (10).
Discussion In each patient studied, measurements made 5 min after interscalene block injection showed a reduction
ANESTH ANALG
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
1992;74352-7
355
Table 2. Percentage Reductions in Pulmonary Function Secondary to Pathological, Surgical, or Interscalene
Block-Induced HemidiaphragmaticParesis Year
Investigators
Etiology of hemidiaphragmatic paresis
1967 1967 1975 1991
Fackler et al. Gould et al. Arborelius et al. Urmey and McDonald
Complete unilateral surgical phrenic nerve section Direct local anesthetic infiltration of phrenic nerve Pathological paralysis of one hemidiaphragm Interscalene block
No. of patients 14 8 17 8
Decrease in FVC 22.7 29 25 27
FVC, forced vital capacity.
in pulmonary function similar in magnitude to published results of studies on patients with complete unilateral phrenic nerve paralysis (Table 2)(6-8). For example, the diminution in FVC was 27.2% in our patients. In a study of 14 patients who underwent complete unilateral phrenic nerve section, Fackler et al. (7) found a mean decrease of 22.7% in vital capacity compared to preoperative values. Similarly, in 17 patients with pathological paralysis of one hemidiaphragm, Arborelius et al. (6) found that vital capacity, FEV,.,, and maximum voluntary ventilation were each reduced by an average of about 25% of predicted values. In a study of direct infiltration of one phrenic nerve with 25 mL of 1% mepivacaine in eight healthy volunteers, Gould et al. (8) found that inspiratory capacity, vital capacity, and total lung capacity were reduced by 27%, 29%, and 20%, respectively. Our comparable findings of mean reductions in FEV,,, and FVC of 26% and 27%, respectively, imply that functionally near complete paralysis of the ipsilateral hemidiaphragm occurred in every patient 5 min after interscalene block injection. These previous studies were presumably done in the upright position, which is less sensitive for diagnosing diaphragmatic paresis because of the weight of the abdominal contents that act to pull the flaccid paretic hemidiaphragm downward (11). Our investigation was done on patients in the supine position, in which the subject‘s abdominal contents push the paretic diaphragm in a cephalad direction and therefore may increase the degree of pulmonary dysfunction by means of a diminution in vital capaaty. We have since found in a separate study of 20 patients that a small further decrement in pulmonary fundion occurred when follow-up measurements were made after the 5min point following interscalene local anesthetic injection (12). We believe that the major mechanism responsible for these measured decreases in spirometric volumes is diminished inspiratory strength resulting from diaphragmatic paresis. This leads to an inability to fully inspire to total lung capacity and a resultant decrease in FVC,as well as in the other volumes and flows that constitute routine pulmonary function testing. Any reduction in maximum inspiratory mus-
cle effort will directly translate to at least an equal diminution in measured expiratory volume. Maximum expiratory flows at any given volume will also decrease, as airflow from the lung is a single-valued function of lung volume. However, in addition to its role as the major muscle of inspiration, the diaphragm may also have some degree of expiratory function (13).To quantify any reduction in expiratory function that occurred during hemidiaphragmatic paresis, data were analyzed by volume-indexing the flow-volume measurements (see Methods). Measuring any true loss in expiratory function is important in increasing our understanding of how this anesthetic affects the patient’s ability to cough perioperatively. The validity of the analysis used depends on there being no decrease in RV. Data from previous investigations showed increases in RV with unilateral hemidiaphragmatic paralysis (6,14). However, although we do not believe that RV decreased in our patients 5 min after interscalene block, we did not use gas dilution or plethysmographic techniques to evaluate this. If, in fact, an increase in RV, as reported previously, occurred at this point, any measured reduction in PEFR that we found by volume-indexing would indeed underestimate the actual reduction. In any event, the patient under interscalene anesthesia has a decreased ability to deep breathe, cough, and clear secretions. The block may also interfere with the patient’s ability to perform effective incentive spirometry in the immediate postoperative period. Reports of altered chest wall mechanics during complete pathological paresis of both hemidiaphragms have been published previously (11,14,15). In these studies, normal and deep inspiration (the active part of normal breathing) were characterized by increased inspiratory upper rib cage expansion and abdominal paradoxical (inward) motion as illustrated in Figure 3. Therefore, in addition to causing large measurable decreases in standard pulmonary function tests, phrenic nerve paresis changes the actual mechanics of respiration. During interscalene block, we found consistent similarly characterized distortions in chest wall mechanics, but with abdominal paradoxical motion onZy on the side of the block.
3-56
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
Figure 3. Illustration of typical changes in chest wall motion measured by magnetometry during interscalene anesthesia. The ipsilateral hemiabdomen no longer expanded, contracting instead with tidal inspiration 10 min after interscalene brachial plexus injection of mepivacaine.
Underlying lung expansion follows regional chest wall expansion. As the lower rib cage has been shown to act in the same manner and move in conjunction with the abdomen (15), unilateral hemidiaphragmatic paresis may lead to decreased lower lung expansion on the side of the block. In fact, decreased regional ventilation has been measured on the affected side in unilateral hemidiaphragmatic paralysis (6). Thus, chest wall distortion during hemidiaphragmatic paresis may lead to regional atelectasis (2). Paralysis or severe weakness of the diaphragm does not by itself lead to respiratory failure in the completely healthy patient (17). However, normal respiration depends on timed coordination between the diaphragm and the other respiratory muscles (9,16,17). The activity of the compensatory muscles (intercostals and accessory) has been shown to decrease during rapid eye movement sleep, and ventilation is reduced during sleep in dogs with acute diaphragmatic paralysis (19). Interscalene block is often done in conjunction with general anesthesia or sedation, and the resulting inhibition of intercostal and accessory muscles during sleep or sedation may be detrimental to lung expansion and gas exchange. For this reason, we administer oxygen to all of our patients receiving this form of regional anesthesia. In patients who are dependent on intact diaphragmatic function and cannot recruit intercostal or accessory muscles to effect, as in ankylosing spondylitis, ventilatory failure may be precipitated by hemidiaphragmatic paresis (20). The fact that large reductions in pulmonary function tests were measured in our two patients with preexisting pulmonary disease has implications for determining anesthetic choice in pa-
ANESTH ANALG 1992;74352-7
tients with moderate or severe obstructive lung disease. Patients unable to tolerate a potential 25% reduction in pulmonary function because of preexisting disease may benefit from another anesthetic technique, either general anesthesia with controlled ventilation or, if possible, axillary block as this regional anesthetic does not paralyze the diaphragm (2). Although infrequent, when a patient does develop dyspnea after interscalene injection, treatment should include reassurance, administration of oxygen, and placement of the patient in the sitting position if possible to optimize diaphragmatic mechanics. If these measures fail, intermittent positive pressure by mask or controlled ventilation should be instituted. There are a few potential weaknesses concerning our study protocol, and data interpretation must be made while keeping these in mind. First, our measurements were made only 5 min after the block. This is a bit premature for complete major nerve blockade to occur, which may take 10-20 min for maximum local anesthetic effect. We have since made sequential measurements on more than 60 patients for a study period of 30 min after interscalene block and have found that only a small further decrement in pulmonary function occurred (12). We were indeed surprised to find that paresis of the phrenic nerve with the measured degree of pulmonary dysfunction occurred so rapidly after interscalene block. We have since found that all pulmonary function changes are essentially complete by 15 min (12). If respiratory insufficiency does occur, it should therefore be seen early, while the patient is under the care of the anesthetist. Second, we used a 45-mL dose of local anesthetic. There is some evidence that when using a single-injection technique, large volumes of local anesthetic produce a more reliable block than smaller volumes (21). We have found in a more recent study that a 20-mL dose produced no significant difference in onset or degree of pulmonary dysfunction when compared in a randomized fashion with a 45-mL dose (12). Finally, interscalene block may also cause paresis of the accessory muscles of inspiration and of the scalene muscles that are classified as inspiratory muscles, and this may have contributed to the pulmonary function decreases measured. In conclusion, interscalene block caused large reductions in standard pulmonary function tests in every patient studied. The reductions in pulmonary function were similar in magnitude to those associated with complete unilateral phrenic nerve paralysis published previously. This anesthetic technique separately reduces inspiratory and expiratory abilities. On the basis of altered regional chest wall expansion, interscalene block may lead to variations in ventilation distribution, regional expiratory flows during
ANESTH ANALG 1992;74:352-7
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
cough, and diminished lung expansion. Each of these factors could lead to postoperative atelectasis or other pulmonary morbidity, especially if long-acting local anesthetic agents or continuous infusions of local anesthetic are used for postoperative analgesia. We thank Dr. Nigel E. Sharrock for his careful review of this manuscript.
References 1 . Winnie AP. Interscalene brachial plexus block. Anesth Analg
1970;49:45546. 2. Urmey WF, Talts KH, Sharrock NE. One hundred percent incidence of hemidiaphragmatic paresis associated with interscalene brachial plexus anesthesia as diagnosed by ultrasonography. Anesth Analg 1991;72:496503. 3. Park GR, Young GB. Demonstration of phrenic nerve paralysis and its subsequent recovery by ultrasound. Intensive Care Med 1981;71454. 4. McCauley RGK, Labib KB. Diaphragmatic paralysis evaluated by phrenic nerve stimulation during fluoroscopy or real-time ultrasound. Radiology 1984;153:33-6. 5. Diament MJ, Boechat MI, Kangerloo H. Real-time sector ultrasound in the evaluation of suspected abnormalities of diaphragmatic motion. J Clin Ultrasound 1985;13:53943. 6. Arborelius M, Lilja 8, Senyk J. Regional and total lung function studies in patients with hemidiaphragmatic paralysis. Respiration 1975;32:253-64. 7. Fackler CD, Perret GE, Ekdell GN. Effect of unilateral phrenic nerve section on lung function. J Appl Physiol 1967;23:923-6. 8. Gould L, Kaplan S, McElhinney AJ, Stone DJ. A method for the production of hemidiaphragmatic paralysis. Its application to the study of lung function in normal man. Am Rev Respir Dis 1967;96812-4.
357
9. DeTroyer A, Estenne M. Coordination between rib cage muscles and diaphragm during quiet breathing in humans. J Appl Physiol 1984;57899-906. 10. Konno K, Mead J. Measurement of the separate volumes of the rib cage and abdomen during breathing. J Appl Physiol 1967; 22~407-22. 11. Kreitzer SM, Feldman NT, Saunders NA, lngram R. Bilateral diaphragmatic paralysis with hypercapnic respiratory failure. Am J Med 1978;65:89-95. 12. Urmey WF, Gloeggler PJ. Hemidiaphragm paresis during interscalene block effects of decreasing local anesthetic injection volume (abstract). Reg Anesth 1991;15(Suppl):35. 13. Agostoni E, Sant’Ambrogio G, Portillo Carrasco, H. Electromyography of the diaphragm in man and transdiaphragmatic pressure. J Appl Physiol 1960;15:1093-7. 14. Lisboa C, Pare PD, Pertuze J, et al. Inspiratory muscle function in unilateral diaphragmatic paralysis. Am Rev Respir Dis 1986;134:48692. 15. Gibson GJ. Diaphragmatic paresis: pathophysiology, clinical features, and investigation. Thorax 1989;44:960-70. 16. Urmey WF, DeTroyer A, Kelly KB, Loring SH. Pleural pressure increases in the zone of apposition of diaphragm to rib cage. J Appl Physiol 1988;65:2207-12. 17. Laroche CM, Carroll N, Moxham J, Green M. Clinical significance of severe isolated diaphragm weakness. Am Rev Respir Dis 1988;138.862-6. 18. Ninane V, Farkas GA, Baer R, DeTroyer A. Mechanism of rib cage inspiratory muscle recruitment in diaphragmatic paralysis. Am Rev Respir Dis 1989;139:146-9. 19. Stradling JR, Kozar LF, Dark J, Kirby T, Andrey SM, Phillipson EA. Effect of acute diaphragm paralysis on ventilation in awake and sleeping dogs. Am Rev Respir Dis 1987;136:633-7. 20. Efthimiou J, McLelland J, Round J, Gibbin HR, Loh L, Spiro SG. Diaphragm paralysis causing ventilatory failure in an adult with rigid spine syndrome. Am Rev Respir Dis 1987;136: 1483-5. 21. Pihlajamaki KK, Lindberg LP. Bupivacaine with and without adrenaline in interscalene brachial plexus blockade. Br J Anaesth 1987;59:1420-4.
MD,
and Marianne McDonald,
BS, RRT
Department of Anesthesiology, The Hospital for Special Surgery, Cornell University Medical College, New York, New York
We studied the effects of unilateral hemidiaphragmatic paresis caused by interscalene brachial plexus block on routine pulmonary function in eight patients. In an additional four patients, we studied changes in chest wall motion during interscalene block anesthesia by chest wall magnetometry. Ipsilatera1 hemidiaphragmatic paresis, as diagnosed by ultrasonography, developed in all patients within 5 min of interscalene injection of 45 mL of 1.5% mepivacaine with added epinephrine and bicarbonate. Large decreases in all pulmonary function variables were measured in every patient. Forced vital capacity and forced expiratory volume at 1 s de-
B
rachial plexus anesthesia by interscalene injection of local anesthetic routinely causes sensory anesthesia of the fourth and fifth, and frequently of the third, cervical nerve. Using radiographic contrast labeling, Winnie (1) demonstrated that this was due to local anesthetic entering the cervical plexus after interscalene brachial plexus injections as small as 10 mL. Local anesthetic entering the cervical plexus at these levels probably results in one-sided diaphragmatic paresis, as the C-3-C-5 nerve roots form the phrenic nerve. In a prior investigation of 13 healthy patients (2), we found that interscalene block resulted in a 100% incidence of ipsilateral hemidiaphragmatic paresis. Diagnosis was made by ultrasonography (3-5), which clearly demonstrated paradoxical motion of the cephalad border of the zone of apposition of diaphragm to rib cage on the ipsilateral side during forceful sniff (a maneuver that rapidly decreases pleural pressure). Previous studies have shown consistent sizeable reductions in pulmonary f unction during pathological Accepted for publication October 21, 1991. Address correspondence to Dr. Urmey, Department of Anesthesiology, The Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021.
352
Anesth Analg 1992;74:352-7
creased 27% 2 4.3% and 26.4% f 6.8%,respectively (P = 0.0001). Peak expiratory and maximum midexpiratory flow rates were also significantly reduced. Interscalene block caused changes in pulmonary function and chest wall mechanical motion that were similar to those published in previous studies on patients with hemidiaphragmatic paresis of pathological or surgical etiology. Interscalene block probably should not be performed in patients who are dependent on intact diaphragmatic function and in those patients unable to tolerate a 25%reduction in pulmonary function. (Anesth Analg 1992;74352-7)
unilateral phrenic nerve paralysis (6,7) as well as during direct local anesthetic blockade of one phrenic nerve (8). We determined that alterations in routine pulmonary function tests and chest wall mechanics occur as the result of unilateral hemidiaphragmatic paresis produced by interscalene block and how these pulmonary function test changes might compare to those previously published during complete unilateral pathological phrenic nerve paralysis. Chest wall magnetometry was used to characterize the physiologic alterations in chest wall motion that occurred during the hemidiaphragmatic paresis that resulted from interscalene block.
Methods After approval from our Institutional Review Board and written informed consent had been obtained, eight consecutive adult patients were studied. Patients were ASA class 1-111 and had been scheduled for upper extremity orthopedic procedures under interscalene brachial plexus anesthesia. Two patients had preexisting pulmonary disease (patient 3, moderate chronic obstructive pulmonary disease and patient 7, asthma). No premedication or sedation was 61992 by the International Anesthesia Research Society 0003-2999/92/$5.00
ANESTH ANALG 1992;74:352-7
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATICPARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
given until all testing was completed. Interscalene block was administered with the needle tip at, or caudad to, cricoid level according to the method of Winnie (1). A single paresthesia technique with a 23-gauge 2.5-cm needle was used. Each patient received 45 mL of 1.5% mepivacaine with 5 pg/mL of epinephrine and 0.05 mEq/mL of sodium bicarbonate. Before anesthesia, real-time ultrasonography (Diasonics DRF-400, 3 - m transducer) ~ of the ipsilateral hemidiaphragm was done to c o n f m normal caudad motion during sniff. After application of noseclips, data were collected during two successive forced vital capacity maneuvers into a computerized bell spirometer (Warren E. Collins, Eagle II, Braintree, Mass.) with coaching by a respiratory therapist. Expiratory and inspiratory flow-volume curves were then generated by forced expiration and inspiration into a Wedge Spirometer (Med Science, St. Louis, Mo.). X-Y flow-volume data were recorded from a dual-trace storage oscilloscope (Tektronix, Beaverton, Ore.) for later analysis. These measurements were repeated 5 min after completion of interscalene local anesthetic injection, by which time all patients had ultrasonographically diagnosed ipsilateral hemidiaphragmatic paresis (paradoxical cephalad hemidiaphragmatic motion during s n q . To determine whether a reduction in expiratory function occurred during hemidiaphragmatic paresis that was not based on a reduction in total lung capacity, data were analyzed by volume-indexing the flow-volume curves. This was done by superimposing the curves at the residual volume (RV) points and comparing the postanesthesia peak expiratory flow rate (PEFR) with the baseline flow rate at the same lung volume, as shown in Figure 1. To analyze the changes in chest wall mechanical motion that occur during interscalene block, an additional four patients were studied using chest wall magnetometry (GMG Scientific, Boston, Mass.). Measurements of anteroposterior diameter changes of the rib cage at midsternum, of the left of the abdomen, and separately of the right half of the abdomen at the umbilical level were made during tidal and deep breathing (Figure2). Data were recorded immediately before and 15min after interscalene block and 10 min after ultrasonographically diagnosed hemidiaphragmatic paresis. X-Y plots of these data were photographed from the dual-trace storage oscilloscope.
Results Ultrasonographic examination revealed that in all eight patients normal ipsilateral hemidiaphragmatic caudad motion changed to paradoxical cephalad motion during forceful sniff within 2 min of interscalene
353
Figure 1. Representative data from a healthy patient comparing baseline (before interscalene block) maximum expiratory flowvolume curve with those measured at 2 and 5 min after interscalene injection of 45 mL of 1.5%mepivacaine with epinephrine. Decreased peak flows at same volumes (referenced to residual volume, RV) constitute evidence that the intact diaphragm normally contributes to forced expiration. TLC represents total lung capacity before interscalene block. See text for discussion.
Rib
Cagr AP Memrter
Figure 2. Representative modified Konnc-Mead diagram is shown for one of the study patients. Vertical axis, ipsilateral and contralateral hemiabdominal anteroposterior (AP) diameter; horizontal axis, rib cage anteroposterior diameter. Tidal breathing loops beginning at functional residual capaaty (FRC) are shown. Ipsilateral hemiabdominal expansion during inspiration changed to contraction (paradoxical motion) during interscalene block in every patient. Contralateral hemiabdominal expansion during inspiration was more preserved. All patients had comparatively more rib cage than abdominal expansion with inspiration during interscalene anesthesia, reflecting recruitment of intercostal and accessory muscles of inspiration to compensate for compromised diaphragmatic inspiratory action. Solid line, before anesthesia; dashed line, during anesthesia.
anesthetic injection. Thus, ipsilateral hemidiaphragmatic paresis developed in 100% of the patients studied, as had been observed in our previous study. Spirometry results for patients 3 and 7 who had preexisting obstructive lung disease showed changes during anesthesia that were similar in value to those measured in the six healthy patients. Table 1 displays the results of routine pulmonary function testing. All eight patients had large reduc-
354
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
ANESTH ANALG 1992;74:352-7
Table 1. Comparison of Pulmonary Function Measurements Before and During Interscalene Block Spirometry Patient No. 1
2 3 4 5 6
7 8
Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change Preblock Postblock % Change
Flowlvolume loop
FVC (L)
FEV1.0 (L)
PEFR (Us)
FEF,,, (Us)
PEFR (Lls)
PIFR (Us)
3.36 2.50 -25.60 3.73 2.45 -34.32 5.63 4.00 -28.95 5.96 4.19 -29.70 4.14 2.94 -28.99 4.52 3.58 -20.80 2.41
3.02 2.20 -27.15 3.32 2.17 -34.64 4.43 3.32 -25.06 4.11 3.04 -26.03 3.09 2.36 -23.62 3.57 2.97 -16.81 1.27 0.80 -37.01 3.98 3.17 -20.35
7.39 6.06 -18.00 6.50 5.25 -19.23 11.80 10.40 -11.86 10.00 7.60 -24.00 5.18 5.12 -1.16 6.99 7.72 10.44 2.37 1.36 -42.62 8.70 7.32 -15.86
3.73 2.52 -32.44 4.22 2.88 -31.75 3.85 3.09 -19.74 2.62 2.03 -22.52 2.48 2.47 -0.40 3.37 3.07 -8.90 0.47 0.41 -12.77 4.14 3.56 -14.01
6.0 5.1 -15.0 5.6 5.5 -1.8 8.0 7.9 -1.3 7.4 6.5 -12.2 5.8 4.5 -22.4 8.1 8.4 3.7 1.3 1.3 0.0 7.0 7.0 0.0
3.8 3.8 0.0 5.9 3.3 -44.9 7.5 7.3 -2.7 4.0 3.9 -2.5 5.0 3.8 -24.0 5.5 5.8 5.5 1.0 1.8 80.0 7.0 6.0 -14.3
1.77 -26.56 4.77 3.70 -22.43
FVC, forced vital capacity; FEV,,,, forced expiratory volume at 1 s; PEFR, peak expiratory flow rate; FEF2s75, forced expiratory flow hom 25% to 75%of the vital capacity; PIFR, peak inspiratory flow rate.
tions in forced vital capacity (FVC) ranging from 21% to 34% within 5 min of interscalene local anesthetic injection. Mean FVC decreased 27.2% t 4.3% (P = 0.0001) for the group. Similarly, forced expiratory volume at 1 s (FEVl.o)decreased between 17% and 37%, with a mean reduction of 26.4% 2 6.8% (P = 0,0001). Postanesthesia peak expiratory flow rate decreased in all but one patient, with a mean reduction of 15.4% t 15.7% (P < 0.05). Maximum midexpiratory flow rate from 25% to 75% of the vital capacity decreased in every patient with a mean diminution of 17.9% t 11.1% (P = 0.05). Volume-indexed PEFR decreased 6.0% t 9.1% (P = 0.05) during interscalene block (Figure 1). Inspiratory flow-volume curve analysis showed that peak inspiratory flow rates declined in five of eight patients and was unchanged in one patient. However, these data were less reproducible (Table l)than expiratory flow and volume data. Using magnetometry, ipsilateral and contralateral hemiabdominal anteroposterior diameters were plotted as separate functions of increasing rib cage anteroposterior diameter during spontaneous inspiration in four patients. A representative tracing from one of the patients is shown in Figure 2. After interscalene block, rib cage motion was more pronounced than abdominal motion in all four patients.
Because of the nature of the measurements, anteroposterior diameter changes by magnetometry may be small during normal or deep breathing as one chest wall compartment may not move at all during breathing. However, the direction of motion and comparison of the chest wall compartments may yield valuable information regarding chest wall muscle function during breathing. In this case, the increase in rib cage activity reflected heightened activity of intercostal and accessory muscles of inspiration (muscles that act on the rib cage and not the abdomen), which are recruited to take over for the compromised diaphragm (9). In addition, paradoxical (inward) ipsilateral hemiabdominal motion during quiet or deep inspiration developed in each patient after receiving interscalene block. By contrast to the ipsilateral side, contralateral hemiabdominal expansion was diminished in three of four patients (and actually increased in the one illustrated patient), but normal directional motion was preserved in all four patients. Thus, the abdomen departed from its usual one degree of freedom of motion (10).
Discussion In each patient studied, measurements made 5 min after interscalene block injection showed a reduction
ANESTH ANALG
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
1992;74352-7
355
Table 2. Percentage Reductions in Pulmonary Function Secondary to Pathological, Surgical, or Interscalene
Block-Induced HemidiaphragmaticParesis Year
Investigators
Etiology of hemidiaphragmatic paresis
1967 1967 1975 1991
Fackler et al. Gould et al. Arborelius et al. Urmey and McDonald
Complete unilateral surgical phrenic nerve section Direct local anesthetic infiltration of phrenic nerve Pathological paralysis of one hemidiaphragm Interscalene block
No. of patients 14 8 17 8
Decrease in FVC 22.7 29 25 27
FVC, forced vital capacity.
in pulmonary function similar in magnitude to published results of studies on patients with complete unilateral phrenic nerve paralysis (Table 2)(6-8). For example, the diminution in FVC was 27.2% in our patients. In a study of 14 patients who underwent complete unilateral phrenic nerve section, Fackler et al. (7) found a mean decrease of 22.7% in vital capacity compared to preoperative values. Similarly, in 17 patients with pathological paralysis of one hemidiaphragm, Arborelius et al. (6) found that vital capacity, FEV,.,, and maximum voluntary ventilation were each reduced by an average of about 25% of predicted values. In a study of direct infiltration of one phrenic nerve with 25 mL of 1% mepivacaine in eight healthy volunteers, Gould et al. (8) found that inspiratory capacity, vital capacity, and total lung capacity were reduced by 27%, 29%, and 20%, respectively. Our comparable findings of mean reductions in FEV,,, and FVC of 26% and 27%, respectively, imply that functionally near complete paralysis of the ipsilateral hemidiaphragm occurred in every patient 5 min after interscalene block injection. These previous studies were presumably done in the upright position, which is less sensitive for diagnosing diaphragmatic paresis because of the weight of the abdominal contents that act to pull the flaccid paretic hemidiaphragm downward (11). Our investigation was done on patients in the supine position, in which the subject‘s abdominal contents push the paretic diaphragm in a cephalad direction and therefore may increase the degree of pulmonary dysfunction by means of a diminution in vital capaaty. We have since found in a separate study of 20 patients that a small further decrement in pulmonary fundion occurred when follow-up measurements were made after the 5min point following interscalene local anesthetic injection (12). We believe that the major mechanism responsible for these measured decreases in spirometric volumes is diminished inspiratory strength resulting from diaphragmatic paresis. This leads to an inability to fully inspire to total lung capacity and a resultant decrease in FVC,as well as in the other volumes and flows that constitute routine pulmonary function testing. Any reduction in maximum inspiratory mus-
cle effort will directly translate to at least an equal diminution in measured expiratory volume. Maximum expiratory flows at any given volume will also decrease, as airflow from the lung is a single-valued function of lung volume. However, in addition to its role as the major muscle of inspiration, the diaphragm may also have some degree of expiratory function (13).To quantify any reduction in expiratory function that occurred during hemidiaphragmatic paresis, data were analyzed by volume-indexing the flow-volume measurements (see Methods). Measuring any true loss in expiratory function is important in increasing our understanding of how this anesthetic affects the patient’s ability to cough perioperatively. The validity of the analysis used depends on there being no decrease in RV. Data from previous investigations showed increases in RV with unilateral hemidiaphragmatic paralysis (6,14). However, although we do not believe that RV decreased in our patients 5 min after interscalene block, we did not use gas dilution or plethysmographic techniques to evaluate this. If, in fact, an increase in RV, as reported previously, occurred at this point, any measured reduction in PEFR that we found by volume-indexing would indeed underestimate the actual reduction. In any event, the patient under interscalene anesthesia has a decreased ability to deep breathe, cough, and clear secretions. The block may also interfere with the patient’s ability to perform effective incentive spirometry in the immediate postoperative period. Reports of altered chest wall mechanics during complete pathological paresis of both hemidiaphragms have been published previously (11,14,15). In these studies, normal and deep inspiration (the active part of normal breathing) were characterized by increased inspiratory upper rib cage expansion and abdominal paradoxical (inward) motion as illustrated in Figure 3. Therefore, in addition to causing large measurable decreases in standard pulmonary function tests, phrenic nerve paresis changes the actual mechanics of respiration. During interscalene block, we found consistent similarly characterized distortions in chest wall mechanics, but with abdominal paradoxical motion onZy on the side of the block.
3-56
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
Figure 3. Illustration of typical changes in chest wall motion measured by magnetometry during interscalene anesthesia. The ipsilateral hemiabdomen no longer expanded, contracting instead with tidal inspiration 10 min after interscalene brachial plexus injection of mepivacaine.
Underlying lung expansion follows regional chest wall expansion. As the lower rib cage has been shown to act in the same manner and move in conjunction with the abdomen (15), unilateral hemidiaphragmatic paresis may lead to decreased lower lung expansion on the side of the block. In fact, decreased regional ventilation has been measured on the affected side in unilateral hemidiaphragmatic paralysis (6). Thus, chest wall distortion during hemidiaphragmatic paresis may lead to regional atelectasis (2). Paralysis or severe weakness of the diaphragm does not by itself lead to respiratory failure in the completely healthy patient (17). However, normal respiration depends on timed coordination between the diaphragm and the other respiratory muscles (9,16,17). The activity of the compensatory muscles (intercostals and accessory) has been shown to decrease during rapid eye movement sleep, and ventilation is reduced during sleep in dogs with acute diaphragmatic paralysis (19). Interscalene block is often done in conjunction with general anesthesia or sedation, and the resulting inhibition of intercostal and accessory muscles during sleep or sedation may be detrimental to lung expansion and gas exchange. For this reason, we administer oxygen to all of our patients receiving this form of regional anesthesia. In patients who are dependent on intact diaphragmatic function and cannot recruit intercostal or accessory muscles to effect, as in ankylosing spondylitis, ventilatory failure may be precipitated by hemidiaphragmatic paresis (20). The fact that large reductions in pulmonary function tests were measured in our two patients with preexisting pulmonary disease has implications for determining anesthetic choice in pa-
ANESTH ANALG 1992;74352-7
tients with moderate or severe obstructive lung disease. Patients unable to tolerate a potential 25% reduction in pulmonary function because of preexisting disease may benefit from another anesthetic technique, either general anesthesia with controlled ventilation or, if possible, axillary block as this regional anesthetic does not paralyze the diaphragm (2). Although infrequent, when a patient does develop dyspnea after interscalene injection, treatment should include reassurance, administration of oxygen, and placement of the patient in the sitting position if possible to optimize diaphragmatic mechanics. If these measures fail, intermittent positive pressure by mask or controlled ventilation should be instituted. There are a few potential weaknesses concerning our study protocol, and data interpretation must be made while keeping these in mind. First, our measurements were made only 5 min after the block. This is a bit premature for complete major nerve blockade to occur, which may take 10-20 min for maximum local anesthetic effect. We have since made sequential measurements on more than 60 patients for a study period of 30 min after interscalene block and have found that only a small further decrement in pulmonary function occurred (12). We were indeed surprised to find that paresis of the phrenic nerve with the measured degree of pulmonary dysfunction occurred so rapidly after interscalene block. We have since found that all pulmonary function changes are essentially complete by 15 min (12). If respiratory insufficiency does occur, it should therefore be seen early, while the patient is under the care of the anesthetist. Second, we used a 45-mL dose of local anesthetic. There is some evidence that when using a single-injection technique, large volumes of local anesthetic produce a more reliable block than smaller volumes (21). We have found in a more recent study that a 20-mL dose produced no significant difference in onset or degree of pulmonary dysfunction when compared in a randomized fashion with a 45-mL dose (12). Finally, interscalene block may also cause paresis of the accessory muscles of inspiration and of the scalene muscles that are classified as inspiratory muscles, and this may have contributed to the pulmonary function decreases measured. In conclusion, interscalene block caused large reductions in standard pulmonary function tests in every patient studied. The reductions in pulmonary function were similar in magnitude to those associated with complete unilateral phrenic nerve paralysis published previously. This anesthetic technique separately reduces inspiratory and expiratory abilities. On the basis of altered regional chest wall expansion, interscalene block may lead to variations in ventilation distribution, regional expiratory flows during
ANESTH ANALG 1992;74:352-7
REGIONAL ANESTHESIA AND PAIN MANAGEMENT URMEY AND McDONALD HEMIDIAPHRAGMATIC PARESIS DURING INTERSCALENE BRACHIAL PLEXUS BLOCK
cough, and diminished lung expansion. Each of these factors could lead to postoperative atelectasis or other pulmonary morbidity, especially if long-acting local anesthetic agents or continuous infusions of local anesthetic are used for postoperative analgesia. We thank Dr. Nigel E. Sharrock for his careful review of this manuscript.
References 1 . Winnie AP. Interscalene brachial plexus block. Anesth Analg
1970;49:45546. 2. Urmey WF, Talts KH, Sharrock NE. One hundred percent incidence of hemidiaphragmatic paresis associated with interscalene brachial plexus anesthesia as diagnosed by ultrasonography. Anesth Analg 1991;72:496503. 3. Park GR, Young GB. Demonstration of phrenic nerve paralysis and its subsequent recovery by ultrasound. Intensive Care Med 1981;71454. 4. McCauley RGK, Labib KB. Diaphragmatic paralysis evaluated by phrenic nerve stimulation during fluoroscopy or real-time ultrasound. Radiology 1984;153:33-6. 5. Diament MJ, Boechat MI, Kangerloo H. Real-time sector ultrasound in the evaluation of suspected abnormalities of diaphragmatic motion. J Clin Ultrasound 1985;13:53943. 6. Arborelius M, Lilja 8, Senyk J. Regional and total lung function studies in patients with hemidiaphragmatic paralysis. Respiration 1975;32:253-64. 7. Fackler CD, Perret GE, Ekdell GN. Effect of unilateral phrenic nerve section on lung function. J Appl Physiol 1967;23:923-6. 8. Gould L, Kaplan S, McElhinney AJ, Stone DJ. A method for the production of hemidiaphragmatic paralysis. Its application to the study of lung function in normal man. Am Rev Respir Dis 1967;96812-4.
357
9. DeTroyer A, Estenne M. Coordination between rib cage muscles and diaphragm during quiet breathing in humans. J Appl Physiol 1984;57899-906. 10. Konno K, Mead J. Measurement of the separate volumes of the rib cage and abdomen during breathing. J Appl Physiol 1967; 22~407-22. 11. Kreitzer SM, Feldman NT, Saunders NA, lngram R. Bilateral diaphragmatic paralysis with hypercapnic respiratory failure. Am J Med 1978;65:89-95. 12. Urmey WF, Gloeggler PJ. Hemidiaphragm paresis during interscalene block effects of decreasing local anesthetic injection volume (abstract). Reg Anesth 1991;15(Suppl):35. 13. Agostoni E, Sant’Ambrogio G, Portillo Carrasco, H. Electromyography of the diaphragm in man and transdiaphragmatic pressure. J Appl Physiol 1960;15:1093-7. 14. Lisboa C, Pare PD, Pertuze J, et al. Inspiratory muscle function in unilateral diaphragmatic paralysis. Am Rev Respir Dis 1986;134:48692. 15. Gibson GJ. Diaphragmatic paresis: pathophysiology, clinical features, and investigation. Thorax 1989;44:960-70. 16. Urmey WF, DeTroyer A, Kelly KB, Loring SH. Pleural pressure increases in the zone of apposition of diaphragm to rib cage. J Appl Physiol 1988;65:2207-12. 17. Laroche CM, Carroll N, Moxham J, Green M. Clinical significance of severe isolated diaphragm weakness. Am Rev Respir Dis 1988;138.862-6. 18. Ninane V, Farkas GA, Baer R, DeTroyer A. Mechanism of rib cage inspiratory muscle recruitment in diaphragmatic paralysis. Am Rev Respir Dis 1989;139:146-9. 19. Stradling JR, Kozar LF, Dark J, Kirby T, Andrey SM, Phillipson EA. Effect of acute diaphragm paralysis on ventilation in awake and sleeping dogs. Am Rev Respir Dis 1987;136:633-7. 20. Efthimiou J, McLelland J, Round J, Gibbin HR, Loh L, Spiro SG. Diaphragm paralysis causing ventilatory failure in an adult with rigid spine syndrome. Am Rev Respir Dis 1987;136: 1483-5. 21. Pihlajamaki KK, Lindberg LP. Bupivacaine with and without adrenaline in interscalene brachial plexus blockade. Br J Anaesth 1987;59:1420-4.
