- Email: [email protected]
Vocal Cord Paralysis After Aortic Surgery
Vocal Cord Paralysis After Aortic Surgery Ralph P. DiLisio, MD,* Michael A. Mazzeffi, MD,† Carol A. Bodian, DrPH,* and Gregory W. Fischer, MD* Objective: The purpose of this study was to investigate variables associated with vocal cord paralysis during complex aortic procedures. Design: A retrospective review. Setting: A tertiary care center. Participants: Four hundred ninety-eight patients who underwent aortic surgery between 2002 and 2007. Methods: Two groups were studied. Group A patients had procedures only involving their aortic root and/or ascending aorta. Group B patients had procedures only involving their aortic arch and/or descending aorta. Results: The incidence of vocal cord paralysis was higher (7.26% v 0.8%) in group B patients (p < 0.0001). Increasing the duration of cardiopulmonary bypass time was associated with an increased risk of vocal cord paralysis and death in both groups A and B (p ⴝ 0.0002 and 0.002, respectively). Additionally, within group B, descending aneurysms emerged as an independent risk factor associated with vocal cord paralysis (p ⴝ 0.03). Length of stay was statistically
significantly longer among group A patients who suffered vocal cord paralysis (p ⴝ 0.017) and trended toward significance in group B patients who suffered vocal cord paralysis (p ⴝ 0.059). The association between tracheostomy and vocal cord paralysis among group A patients reached statistical significance (p ⴝ 0.007) and trended toward significance in group B patients (p ⴝ 0.057). Conclusions: Increasing duration of cardiopulmonary bypass time was associated with a higher risk of vocal cord paralysis in patients undergoing aortic surgery. Additionally, within group B patients, descending aortic aneurysm was an independent risk factor associated with vocal cord paralysis. Most importantly, vocal cord paralysis appeared to have an association between an increased length of stay and tracheostomy among a select group of patients undergoing aortic surgery. © 2013 Elsevier Inc. All rights reserved.
W
2007, were identified using the institutional Anesthesia Information Management System Aims (CompuRecord; Philips, Andover, MA). All patient and perioperative anesthetic data were obtained from this system. Surgical procedures included all types of thoracic aortic surgery including aneurysms, dissections, and infections. For analysis, 2 groups were created based on the anatomic location of the thoracic aneurysm. Patients who had aneurysms limited to the aortic root and/or the ascending aorta were categorized as group A. Patients who had aneurysms only involving the aortic arch and/or the descending thoracic aorta were classified as group B. This division was necessary because the patient characteristics differed between the 2 groups (Table 1), and analyzing all patients together could mask different associations. All patients who underwent thoracic aortic surgery at the authors’ institution received a standard anesthetic technique with central venous and peripheral artery cannulation. Pulmonary artery catheters and transesophageal echocardiography were used in all cases. In all descending aortic aneurysm procedures, a lumbar cerebrospinal fluid drain was inserted. For cases performed via midline sternotomy, the trachea was intubated with a standard high-volume, low-pressure, cuffed, single-lumen endotracheal tube (7.5 mm for females and 8.0 mm for males). Anesthesia was induced with etomidate, midazolam, fentanyl, and an intermediate-acting neuromuscular blocking agent. Anesthesia was maintained with a balanced technique of isoflurane and fentanyl. For cases performed via thoracotomy, a double-lumen endotracheal tube was inserted (37-39 F for females and 39-41 F for males). In select patients in whom neuromonitoring of the spinal cord occurred, volatile agents and neuromuscular blocking drugs were avoided. In these cases, anesthesia was maintained with propofol and fentanyl infusions. Temperature monitoring was performed by the insertion of esophageal and Foley catheter temperature probes. The patients were mechanically ventilated for the duration of the procedure with minute ventilation necessary to achieve normocapnia as measured by capnography. Cardiopulmonary bypass (CPB) was necessary in all cases, and all but 1 patient underwent a period of deep hypothermic circulatory arrest because this was the practice of the thoracic aortic surgeons at the authors’ institution during this time period. After completion of the procedure, all patients were transported intubated to the intensive care unit (ICU). Weaning from the ventilator occurred in compliance with the authors’ institutional ICU protocol. After extubation, if a patient showed signs consistent with vocal cord
ITHIN THE REALM OF cardiac surgery, procedures involving the thoracic aorta are considered some of the most complex surgeries performed and are associated with a number of postoperative complications. Injury to the recurrent laryngeal nerve leading to vocal cord paralysis (VCP) is a complication that has emerged in the literature within the last 5 to 10 years as an independent risk factor for a number of serious adverse outcomes.1 The incidence of VCP after cardiac surgical procedures ranges from 2% to 32%.2,3 Prior investigations into this topic either included all types of cardiac procedures or involved a small number of patients undergoing aortic surgery. The main objective of the current study was to estimate the incidence of VCP after thoracic aortic surgery and to explore the relationship among VCP, tracheostomy, and hospital length of stay (LOS). Additionally, the authors sought to investigate perioperative factors associated with VCP among patients who underwent thoracic aortic surgery. METHODS The Institutional Review Board at the authors’ institution approved the study protocol in September 2011. A total of 498 adult patients who had undergone thoracic aortic surgery from July 1, 2002, to December 30,
From the *Department of Anesthesiology, The Mount Sinai Medical Center, New York, NY; and †Department of Cardiothoracic Anesthesiology and Critical Care, Emory University Hospital, Atlanta, GA. Presented at the 30th Annual Symposium Clinical Update in Anesthesiology, Surgery, and Perioperative Medicine, January 15–20, 2012, Rio Grande, Puerto Rico and this was also presented at The American Society of Anesthesiologists Annual Meeting, October 14, 2012, Washington, DC. Address reprint requests to Ralph DiLisio, MD, Department of Anesthesiology, The Mount Sinai Medical Center, Box 1010, One Gustave L. Levy Place, New York, NY 10029. E-mail: ralph.dilisio@ mountsinai.org © 2013 Elsevier Inc. All rights reserved. 1053-0770/2703-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2012.09.011 522
KEY WORDS: length of stay, thoracic aortic aneurysm, tracheostomy, vocal cord paralysis
Journal of Cardiothoracic and Vascular Anesthesia, Vol 27, No 3 (June), 2013: pp 522-527
VOCAL CORD PARALYSIS
523
Table 1. Demographic and Perioperative Data of Patients Undergoing Aortic Surgery
Demographic Variables
Group A: Aneurysms Involving the Aortic Root and/or Ascending Aorta (n ⫽ 250)
Group B: Aneurysms Involving Aortic Arch and/or Descending Aorta (n ⫽ 248)
Age (y) Males (%) Emergency (%) Sternotomy v thoracotomy (%) Preoperative infection in surgical site (%) Reoperation (%) Descending aortic aneurysm (%)* CPB duration (min) DHCA duration (min)† Esophageal temperature duration below 20°C (min)
61.5 (15, 89) 182 (73) 22 (9) 250 (100) 7 (3) 48 (19) 0 (0) 250 (112, 631) 20 (0, 54) 99 (0, 359)
65 (15, 87) 151 (61) 24 (10) 189 (76) 5 (2) 102 (41) 110 (44) 228 (105, 489) 30 (0, 59) 126 (0, 304)
p Value
0.007 0.005 0.74 ⬍0.0001 0.77 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001
NOTE. Values are median (minimum, maximum) or number (%). Abbreviations: CPB, cardiopulmonary bypass. *Descending aortic aneuryms: aneurysms beginning distal to the left subclavian and extending distally. †Deep hypothermic circulatory arrest.
paralysis or paresis, an otolaryngology consult was obtained. An otolaryngologist then confirmed each diagnosis by fiberoptic laryngeal examination and direct laryngoscopy. The vocal cords of the patients who died were not examined. These patients were either never extubated and thus could not exhibit signs of vocal cord paralysis or died secondary to more significant medical conditions and did not undergo postmortem examinations. The authors chose to include these patients throughout the analysis not only to avoid possible biased estimates but also to provide a complete picture of the results and to show an interesting parallel among patients who died and those who suffered vocal cord paralysis. Because the vocal cord status of patients who died was not known, each patient’s outcome was classified as known VCP, death, or neither. The initial analysis looked at associations among this set of outcomes and demographic and perioperative factors. It became apparent that the strongest association was with the location of the surgical procedure. As seen in Tables 1 and 2, the group of patients who had aneurysms involving the aortic arch and/or descending aorta (group B) were vastly different from those who had procedures limited to the aortic root or ascending aorta (group A) in terms of both the set of outcomes and perioperative and demographic variables. These comparisons were tested using Wilcoxon tests or chi-square tests as appropriate. Because of the significant differences noted, subsequent analyses were conducted separately for the 2 groups of patients. Nominal logistic regression, implemented with SAS PROC CATMOD (SAS Institute Inc., Cary, NC), was used to identify factors associated with the set of outcomes in each group. All factors shown in Table 1 were considered. LOS was compared among the 3 outcomes using Kruskal-Wallis tests with exact Monte Carlo p values implemented with StatXact4 (Cytel Software Corporation, Cambridge, MA) for
Table 2. Clinical Outcomes Within Group A and Group B
Clinical Outcome
Group A: Aneurysms Involving Group B: Aneurysms Involving the Aortic Arch and/or the Aortic Root and/or Descending Aorta Ascending Aorta (n ⫽ 248) (n ⫽ 250)
Vocal cord paralysis (%) Death (%) Neither (%) NOTE: p ⬍ .0001.
2 (0.8) 7 (2.8)
18 (7.26) 20 (8.06)
241 (96.4)
210 (84.68)
Windows. Exact chi-square tests were used to compare the incidence of postoperative tracheostomy among the 3 outcomes. RESULTS
All initial demographic and perioperative data for groups A and B are summarized in Table 1. The incidence of the set of outcomes (ie, VCP, inpatient death, or neither) among group A patients was compared with that among group B (Table 2). This analysis showed a statistically significantly higher incidence of both VCP and death among patients who had aneurysms involving their aortic arch and/or descending aorta (group B). Of the 20 cases of VCP, 19 involved the left vocal cord. The remaining one was a case of bilateral VCP. Further analysis then was performed to determine which of the demographic and operative variables listed in Table 1 were associated with the 3 outcomes. Among the patients who had aneurysms limited to the aortic root and/or ascending aorta (group A), an increasing duration of CPB time emerged as a significant factor associated with the risk of VCP and death (p ⫽ 0.0002) (Table 3). Within group B (ie, procedures involving the aortic arch and/or descending aorta), 2 factors (ie, the duration of CPB and descending aneurysm) emerged as statistically significant independent variables associated with VCP and death (CPB time, p ⫽ 0.002; descending aneurysm, p ⫽ 0.03; Fig 1). None of the additional variables examined approached statistical significance within either group of patients.
Table 3. Vocal Cord Paralysis and Death by the Duration of CPB Time Among Group A Patients Duration of CPB Time (min)
ⱕ200 200-240 240-300 ⬎300
Vocal Cord Paralysis (%)
Death (%)
Neither (%)
0 (0) 0 (0) 1 (1.11) 1 (1.85)
0 (0) 2 (2.78) 0 (0) 5 (9.26)
34 (100) 70 (97.2) 89 (98.8) 48 (88.8)
NOTE. p ⫽ 0.0002. Group A: aneurysms involving the aortic root and/or ascending aorta. Abbreviation: CPB, cardiopulmonary bypass.
524
DILISIO ET AL
Fig 1. Independent risk factors (ie, cardiopulmonary bypass time and descending aortic aneurysm) associated with clinical outcomes among patients within group B. Vocal cord paralysis (), death ( ), and neither ( ). Cardiopulmonary bypass time: p ⴝ 0.002, descending aortic aneurysm: p ⴝ 0.03. Group B: aneurysms involving the aortic arch and/or descending aorta.
To study the postoperative impact of VCP, the authors compared the LOS of the 3 outcome groups (ie, VCP, death, and neither) within each surgical group (Fig 2). Within group A, the median (minimum, maximum) LOS for patients who had VCP was 20.5 days (13, 28 days). The median (minimum, maximum) LOS for the patients who died was 12 days (0, 182 days). The survivors without vocal cord paralysis (VCP) had a median (minimum, maximum) LOS of only 8 days (4, 114 days). Among the group B patients, the median (minimum, maximum) LOS for the VCP group also was 20.5 days (7, 112 days). The median (minimum, maximum) LOS for the patients who died was 9.5 days (1, 101 days). The survivors without VCP had a median (minimum, maximum) LOS of 12 days (4, 182 days). The LOS differences were statistically significant (p ⫽ 0 .017) for group A patients but failed to reach statistical significance in group B patients (p ⫽ 0.059). Among the group A patients, the number (%) with a tracheostomy was 1 (50%), 1 (14.3%), and 4 (1.7%) for VCP, death, and neither, respectively (p ⫽ 0.007). The corresponding numbers for group B patients were 4 (22.2%), 4 (20.0%), and 18 (8.6%), respectively (p ⫽ 0.057, Fig 3). DISCUSSION
The authors have shown a statistically significant association between the increased length of hospital stay and tracheostomy among patients who are diagnosed with vocal cord paralysis after procedures involving their aortic root and/or ascending aorta. Although this association was not statistically significant among patients who suffer the same complication after procedures involving their aortic arch and/or descending aorta, it
Fig 2. The median postoperative LOS by clinical outcomes for groups A and B. Vocal cord paralysis (), death ( ), and neither ( ). Group A: p ⴝ 0.017; group B: p ⴝ 0.059. Group A: aneurysms involving the aortic root and/or ascending aorta. Group B: aneurysms involving the aortic arch and/or descending aorta.
clearly trended in the same direction (p ⫽ 0.059 for LOS and p ⫽ 0.057 for tracheostomy). Among all patients undergoing thoracic aortic surgery, increasing the duration of CPB time was associated with an increased risk of vocal cord paralysis. Independent of the duration of the CPB time, procedures performed on the descending thoracic aorta were also associated with an increased risk of vocal cord paralysis. This is the largest study to date looking specifically at VCP in adults after thoracic aortic surgery. The incidence of VCP was found to be the highest (7.26%) in patients who presented for the repair of aneurysms that involved their aortic arch and/or descending aorta. This percentage was similar to other studies that have reported an incidence of near 9% in procedures at this location.4,5 Other studies that focused primarily on coronary artery bypass graft procedures without the involvement of the ascending aorta reported a risk near 2%.2,6 This is similar to the group A patients despite the fact that additionally all ascending aortic repairs were included in group A. Clearly, the anatomic course of the left recurrent laryngeal nerve as it branches from the vagus nerve at the level of the transverse aortic arch and loops under it, posterior to the ligamentum arteriosum, to ascend up the thorax places it at risk during thoracic aortic surgical procedures. The fact that nearly all cases of VCP in this study involved only the left vocal cord further supports the theory that the nerve was damaged during the surgical procedure itself. However, there are other causes of VCP after surgery. One retrospective study of 31,241 patients reported that 0.08% of patients suffered from VCP simply after tracheal intubation.7 Although this retrospective study did show an increased asso-
VOCAL CORD PARALYSIS
525
Fig 3. The incidence of tracheostomy ( ) and no tracheostomy ( ) within clinical outcomes for groups A and B. Group A: p ⴝ 0.007. Group A: aneurysms involving the aortic root and/or ascending aorta. Group B: p ⴝ 0.057. Group B: aneurysms involving the aortic arch and/or descending aorta.
ciation with prolonged intubation time and increasing age, other prospective investigations have disputed this finding.8 Increased pressure from an inflated endotracheal tube cuff transmitted to the tracheal mucosa near where the external branch of the recurrent laryngeal nerve enters the trachea to innervate the vocal cords is thought to be the mechanism behind intubation-related injury. A condition sometimes mistaken for VCP is arytenoid cartilage dislocation. Rubin et al9 reported that this is an extremely rare complication of traumatic intubation and that it is distinguished from true VCP with strobovideolaryngoscopy. The present authors believe that because all 18 cases of VCP diagnosed in patients who had procedures involving the aortic arch and/or descending aorta were involving the left vocal cord that the respective nerve was injured during the surgical repair. Although damage to the recurrent laryngeal nerve during the surgical procedure is the most likely culprit, it is interesting to note that none of the dictated operative reports mentioned this complication. One investigation was able to show that if the operating surgeon had the impression that the recurrent laryngeal nerve possibly was damaged during surgery that it was in fact associated with a higher probability of vocal cord immobility.10 This may be evidence enough that instead of relying on the operative note, it would be worthwhile to simply ask the operating surgeon if they believe there was a chance that the recurrent laryngeal nerve was damaged during surgery to anticipate potential associated complications.
The finding that the risk of VCP increased with an increasing duration of CPB time in both groups is not surprising. Two other studies have shown that the risk of VCP increased with an increasing duration of both CPB time and total operative time.1,11 This may be a marker for a more complex dissection that makes injury to the nerve more likely. Alternatively, lengthy procedures and increasing the duration of CPB also may be a marker for a longer period during which patients are exposed to colder temperatures, which is another known risk factor for nerve injury. This type of nerve injury often is limited to the immediate time after rewarming and not likely a contributor to permanent damage unless hypothermia is severe and prolonged.12 However, there are case reports of injury in this manner to both the phrenic and recurrent laryngeal nerves producing diaphragmatic and VCP, respectively, lasting nearly 1 year.13 The injury was attributed to a prolonged exposure to ice/slush mix used for myocardial preservation. In the present investigation, no attempt was made to address other potential causes of nerve damage. If hypothermia was a significant source of permanent nerve injury, clinicians would expect to see an association between the incidence of VCP and the duration of an esophageal temperature time below 20°C or a duration of deep hypothermic circulatory arrest. The present authors did not find such associations in their analyses. The present results showed a strong association between the location of the procedure and the risk of VCP. Specifically, in the present study, the risk of experiencing VCP in the group of
526
DILISIO ET AL
patients with aortic arch and/or descending aortic aneurysm repairs (group B) was at least 7.3% compared with only 0.8% for the group A patients, who had aneurysms involving the aortic root and ascending aorta. Classifying the group B patients further in terms of whether the aortic arch or descending thoracic aorta was involved in the actual surgical procedure provides a deeper understanding of how the anatomic location of the pathology influences the risk of VCP. Because of the location of the recurrent laryngeal nerve and the aortic arch, it was not unexpected to find that the risk associated with procedures involving descending aortic aneurysms was significantly greater than with the ascending aorta and/or aortic arch (p ⬍ 0.03). Other reports also have shown that the incidence of postoperative left-sided vocal cord paralysis is greater in patients who undergo repair limited to the descending aorta or distal arch aneurysms.1,11 Ohta and Mori,14 in a series examining 69 patients who underwent surgery on their descending thoracic aorta, found that chronic dilation at the left subclavian artery and surgical anastomosis occurring proximal to the left subclavian artery were independent risk factors for postoperative VCP (odds ratios ⫽ 8.67 and 17.7, respectively). It is quite possible that within the 110 descending aortic aneurysms in the present series the surgeon performed the anastomosis proximal to the left subclavian artery because of the immediacy of the aneurysmal formation following the distal edge of this artery. The cumulative weight of the previously mentioned evidence showed that VCP in the postoperative period was the highest when the pathology and surgical site were located at or near the origin of the left subclavian artery. The association between VCP and a greater length of hospital stay was similar to what other investigators have found.1 This association reached statistical significance among patients who had aneurysms involving the aortic root and/or ascending aorta and approached statistical significance in the group of patients who had procedures that involved the aortic arch and/or descending aorta. Although the additional cost incurred per hospital day is difficult to estimate, clearly a longer LOS further strains an already overburdened health care system.15 The dramatic association between VCP and tracheostomy is cause for concern. The present study showed a 25% incidence of tracheostomy among all patients who had VCP compared with the 5% incidence in the survivors without VCP. Ohta et al11 reported a 22.5% incidence of tracheostomy or reintubation in their cohort of 40 patients who suffered from VCP. Although the present study did not examine additional complications such as swallowing difficulty, decreased cough reflex, and aspiration pneumonia that potentially are associated with VCP leading to tracheostomy, there are multiple studies in both the
pediatric and adult cardiothoracic literature confirming this association.1,16,17 The aryepiglottic muscle, which receives innervation from the recurrent laryngeal nerve, constricts during swallowing, protecting the entrance to the larynx.18 Damage to the recurrent laryngeal nerve exposes the larynx during swallowing, increasing the risk of aspiration. There is evidence that the early treatment of known VCP by injection laryngoplasty reduces the risk of developing aspiration pneumonia.19 A majority of the patients who suffered VCP in the present study did have injection laryngoplasty performed; however, it is unknown if this procedure averted any cases of aspiration pneumonia because these data were unavailable. Aspiration also has been shown to be a risk factor for tracheostomy in a large prospective cohort study of 552 intensive care patients.20 Tracheostomy has been associated with declines in physical and mental functioning from the patient perspective and is viewed widely as a setback in the recovery period.21 This study was limited by the fact it was a retrospective study. Also, the vocal cord status of the 27 patients who died was not confirmed. These data were very difficult to discover because many patients died before they were extubated, and could not be examined for vocal cord mobility. The incidence of VCP in the present cohort was likely underestimated because asymptomatic patients did not undergo evaluation by an otolaryngologist. One study that looked prospectively at the incidence of VCP in children undergoing cardiothoracic surgery showed that 14% of asymptomatic patients suffered from VCP.22 Additionally, some perioperative variables, such as aneurysm size and surgical technique, which may have an association with VCP, were not examined. Despite these limitations, the authors believe these results are significant in light of the limited studies available regarding this topic. Once thought of as a relatively minor complication in comparison with other more serious outcomes such as renal failure, paralysis, or sepsis in this surgical population, VCP may portend another important prognosis regarding tracheostomy and increased LOS. One study of patients undergoing esophagectomy found unilateral VCP in 34% of patients. However, this outcome was not associated with a longer ICU stay or an increase in pulmonary complications.23 The present authors have shown that the same injury to the recurrent laryngeal nerve in thoracic aortic surgical patients does have an association with greater morbidity. ACKNOWLEDGMENT The authors would like to thank John Essad for his help with table and graph editing.
REFERENCES 1. Itagaki T, Kikura M, Sato S: Incidence and risk factors of postoperative vocal cord paralysis in 987 patients after cardiovascular surgery. Ann Thorac Surg 83:2147-2152, 2007 2. Shafei H, El-Kholy A, Azmy S, et al: Vocal cord dysfunction after cardiac surgery: an overlooked complication. Eur J Cardiothorac Surg 11:564-566, 1997 3. Ishimoto S, Ito K, Toyama M, et al: Vocal cord paralysis after surgery for thoracic aortic aneurysm. Chest 121:1911-1915, 2002
4. Ishii K, Adachi H, Tsubaki K, et al: Evaluation of recurrent nerve paralysis due to thoracic aortic aneurysm and aneurysm repair. Laryngoscope 114:2176-2181, 2004 5. Bhaskar J, Foo J, Sharma A. Clamp-and-sew technique for traumatic injuries of the aorta: 20-year experience. Asian Cardiovasc Thorac Ann 18:161-165, 2010 6. Hamdan A, Moukarbel R, Farhat F, et al: Vocal cord paralysis after open-heart surgery. Eur J Cardiothorac Surg 21:671-674, 2002
VOCAL CORD PARALYSIS
7. Kikura M, Suzuki K, Itagaki T, et al: Age and comorbidity as risk factors for vocal cord paralysis associated with tracheal intubation. Br J Anaesth 98:524-530, 2007 8. Colton J, Noordzij J, Murgia B, et al: Laryngeal injury from prolonged intubation: a prospective analysis of contributing factors. Laryngoscope 121:596-600, 2011 9. Rubin A, Hawkshaw M, Moyer C, et al: Arytenoid cartilage dislocation: A 20-year experience. J Voice 19:687-701, 1997 10. Carpes L, Kozak F, Leglanc J, et al: Assessment of vocal fold mobility before and after cardiothoracic surgery in children. Arch Otolaryngol Head Neck Surg 137:571-575, 2011 11. Ohta N, Kuratani T, Hagihira S, et al: Vocal cord paralysis after aortic arch surgery: predictors and clinical outcome. J Vasc Surg 43:721-728, 2006 12. Stecker M, Baylor K: Peripheral nerve at extreme low temperatures 1: effects of temperature on the action potential. Crybiology 59:1-11, 2009 13. Tewari P, Aggarwal S: Combined left-sided recurrent laryngeal and phrenic nerve palsy after coronary artery operation. Ann Thorac Surg 61:1721-1723, 1996 14. Ohta N, Mori T: Vocal cord paralysis after surgery to the descending thoracic aorta via left posterolateral thoracotomy. Ann Vasc Surg 21:761-766, 2007 15. Seidel J, Whiting P, Edbrooke D: The costs of intensive care. Continuing Educ Anaesth Crit Care Pain 6:160-163, 2006
527
16. Sachdeva R, Hussain E, Moss MM, et al: Vocal cord dysfunction and feeding difficulties after pediatric cardiovascular surgery. J Pediatr 151:312-315, 315.e1-2, 2007 17. Benjamin J, Smith B, Cotton M, et al: Long-term morbidities associated with vocal cord paralysis after surgical closure of a patent ductus arteriosus in extremely low birth weight infants. J Perinatol 30:408-413, 2010 18. Thach B: Recurrent laryngeal nerve injury during patent ductus ligation: Can this common complication be reduced? J Perinatol 30: 371-372, 2010 19. Graboyes E, Bradley J, Meyers B, et al: Efficacy and safety of acute injection laryngoplasty for vocal cord paralysis following thoracic surgery. Laryngoscope 212:2406-2410, 2011 20. Kollef M, Ahrens T, Shannon W: Clinical predictors and outcomes for patients requiring tracheostomy in the intensive care unit. Crit Care Med 27:1714-1720, 1999 21. Hashmi N, Ransom E, Nardone H, et al: Quality of life and self-image in patients undergoing tracheostomy. Laryngoscope 120: S196, 2010 22. Smith M, King J, Elsherif A, et al: Should all newborns who undergo patent ductus arteriosus ligation be examined for vocal fold mobility? Laryngoscope 119:1606-1609, 2009 23. Johnson P, Kanegoanker G, Bates T: Indirect laryngoscopic evaluation of vocal cord function in patients undergoing transhiatal esophagectomy. J Am Coll Surg 178:605-608, 1994
significantly longer among group A patients who suffered vocal cord paralysis (p ⴝ 0.017) and trended toward significance in group B patients who suffered vocal cord paralysis (p ⴝ 0.059). The association between tracheostomy and vocal cord paralysis among group A patients reached statistical significance (p ⴝ 0.007) and trended toward significance in group B patients (p ⴝ 0.057). Conclusions: Increasing duration of cardiopulmonary bypass time was associated with a higher risk of vocal cord paralysis in patients undergoing aortic surgery. Additionally, within group B patients, descending aortic aneurysm was an independent risk factor associated with vocal cord paralysis. Most importantly, vocal cord paralysis appeared to have an association between an increased length of stay and tracheostomy among a select group of patients undergoing aortic surgery. © 2013 Elsevier Inc. All rights reserved.
W
2007, were identified using the institutional Anesthesia Information Management System Aims (CompuRecord; Philips, Andover, MA). All patient and perioperative anesthetic data were obtained from this system. Surgical procedures included all types of thoracic aortic surgery including aneurysms, dissections, and infections. For analysis, 2 groups were created based on the anatomic location of the thoracic aneurysm. Patients who had aneurysms limited to the aortic root and/or the ascending aorta were categorized as group A. Patients who had aneurysms only involving the aortic arch and/or the descending thoracic aorta were classified as group B. This division was necessary because the patient characteristics differed between the 2 groups (Table 1), and analyzing all patients together could mask different associations. All patients who underwent thoracic aortic surgery at the authors’ institution received a standard anesthetic technique with central venous and peripheral artery cannulation. Pulmonary artery catheters and transesophageal echocardiography were used in all cases. In all descending aortic aneurysm procedures, a lumbar cerebrospinal fluid drain was inserted. For cases performed via midline sternotomy, the trachea was intubated with a standard high-volume, low-pressure, cuffed, single-lumen endotracheal tube (7.5 mm for females and 8.0 mm for males). Anesthesia was induced with etomidate, midazolam, fentanyl, and an intermediate-acting neuromuscular blocking agent. Anesthesia was maintained with a balanced technique of isoflurane and fentanyl. For cases performed via thoracotomy, a double-lumen endotracheal tube was inserted (37-39 F for females and 39-41 F for males). In select patients in whom neuromonitoring of the spinal cord occurred, volatile agents and neuromuscular blocking drugs were avoided. In these cases, anesthesia was maintained with propofol and fentanyl infusions. Temperature monitoring was performed by the insertion of esophageal and Foley catheter temperature probes. The patients were mechanically ventilated for the duration of the procedure with minute ventilation necessary to achieve normocapnia as measured by capnography. Cardiopulmonary bypass (CPB) was necessary in all cases, and all but 1 patient underwent a period of deep hypothermic circulatory arrest because this was the practice of the thoracic aortic surgeons at the authors’ institution during this time period. After completion of the procedure, all patients were transported intubated to the intensive care unit (ICU). Weaning from the ventilator occurred in compliance with the authors’ institutional ICU protocol. After extubation, if a patient showed signs consistent with vocal cord
ITHIN THE REALM OF cardiac surgery, procedures involving the thoracic aorta are considered some of the most complex surgeries performed and are associated with a number of postoperative complications. Injury to the recurrent laryngeal nerve leading to vocal cord paralysis (VCP) is a complication that has emerged in the literature within the last 5 to 10 years as an independent risk factor for a number of serious adverse outcomes.1 The incidence of VCP after cardiac surgical procedures ranges from 2% to 32%.2,3 Prior investigations into this topic either included all types of cardiac procedures or involved a small number of patients undergoing aortic surgery. The main objective of the current study was to estimate the incidence of VCP after thoracic aortic surgery and to explore the relationship among VCP, tracheostomy, and hospital length of stay (LOS). Additionally, the authors sought to investigate perioperative factors associated with VCP among patients who underwent thoracic aortic surgery. METHODS The Institutional Review Board at the authors’ institution approved the study protocol in September 2011. A total of 498 adult patients who had undergone thoracic aortic surgery from July 1, 2002, to December 30,
From the *Department of Anesthesiology, The Mount Sinai Medical Center, New York, NY; and †Department of Cardiothoracic Anesthesiology and Critical Care, Emory University Hospital, Atlanta, GA. Presented at the 30th Annual Symposium Clinical Update in Anesthesiology, Surgery, and Perioperative Medicine, January 15–20, 2012, Rio Grande, Puerto Rico and this was also presented at The American Society of Anesthesiologists Annual Meeting, October 14, 2012, Washington, DC. Address reprint requests to Ralph DiLisio, MD, Department of Anesthesiology, The Mount Sinai Medical Center, Box 1010, One Gustave L. Levy Place, New York, NY 10029. E-mail: ralph.dilisio@ mountsinai.org © 2013 Elsevier Inc. All rights reserved. 1053-0770/2703-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2012.09.011 522
KEY WORDS: length of stay, thoracic aortic aneurysm, tracheostomy, vocal cord paralysis
Journal of Cardiothoracic and Vascular Anesthesia, Vol 27, No 3 (June), 2013: pp 522-527
VOCAL CORD PARALYSIS
523
Table 1. Demographic and Perioperative Data of Patients Undergoing Aortic Surgery
Demographic Variables
Group A: Aneurysms Involving the Aortic Root and/or Ascending Aorta (n ⫽ 250)
Group B: Aneurysms Involving Aortic Arch and/or Descending Aorta (n ⫽ 248)
Age (y) Males (%) Emergency (%) Sternotomy v thoracotomy (%) Preoperative infection in surgical site (%) Reoperation (%) Descending aortic aneurysm (%)* CPB duration (min) DHCA duration (min)† Esophageal temperature duration below 20°C (min)
61.5 (15, 89) 182 (73) 22 (9) 250 (100) 7 (3) 48 (19) 0 (0) 250 (112, 631) 20 (0, 54) 99 (0, 359)
65 (15, 87) 151 (61) 24 (10) 189 (76) 5 (2) 102 (41) 110 (44) 228 (105, 489) 30 (0, 59) 126 (0, 304)
p Value
0.007 0.005 0.74 ⬍0.0001 0.77 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001
NOTE. Values are median (minimum, maximum) or number (%). Abbreviations: CPB, cardiopulmonary bypass. *Descending aortic aneuryms: aneurysms beginning distal to the left subclavian and extending distally. †Deep hypothermic circulatory arrest.
paralysis or paresis, an otolaryngology consult was obtained. An otolaryngologist then confirmed each diagnosis by fiberoptic laryngeal examination and direct laryngoscopy. The vocal cords of the patients who died were not examined. These patients were either never extubated and thus could not exhibit signs of vocal cord paralysis or died secondary to more significant medical conditions and did not undergo postmortem examinations. The authors chose to include these patients throughout the analysis not only to avoid possible biased estimates but also to provide a complete picture of the results and to show an interesting parallel among patients who died and those who suffered vocal cord paralysis. Because the vocal cord status of patients who died was not known, each patient’s outcome was classified as known VCP, death, or neither. The initial analysis looked at associations among this set of outcomes and demographic and perioperative factors. It became apparent that the strongest association was with the location of the surgical procedure. As seen in Tables 1 and 2, the group of patients who had aneurysms involving the aortic arch and/or descending aorta (group B) were vastly different from those who had procedures limited to the aortic root or ascending aorta (group A) in terms of both the set of outcomes and perioperative and demographic variables. These comparisons were tested using Wilcoxon tests or chi-square tests as appropriate. Because of the significant differences noted, subsequent analyses were conducted separately for the 2 groups of patients. Nominal logistic regression, implemented with SAS PROC CATMOD (SAS Institute Inc., Cary, NC), was used to identify factors associated with the set of outcomes in each group. All factors shown in Table 1 were considered. LOS was compared among the 3 outcomes using Kruskal-Wallis tests with exact Monte Carlo p values implemented with StatXact4 (Cytel Software Corporation, Cambridge, MA) for
Table 2. Clinical Outcomes Within Group A and Group B
Clinical Outcome
Group A: Aneurysms Involving Group B: Aneurysms Involving the Aortic Arch and/or the Aortic Root and/or Descending Aorta Ascending Aorta (n ⫽ 248) (n ⫽ 250)
Vocal cord paralysis (%) Death (%) Neither (%) NOTE: p ⬍ .0001.
2 (0.8) 7 (2.8)
18 (7.26) 20 (8.06)
241 (96.4)
210 (84.68)
Windows. Exact chi-square tests were used to compare the incidence of postoperative tracheostomy among the 3 outcomes. RESULTS
All initial demographic and perioperative data for groups A and B are summarized in Table 1. The incidence of the set of outcomes (ie, VCP, inpatient death, or neither) among group A patients was compared with that among group B (Table 2). This analysis showed a statistically significantly higher incidence of both VCP and death among patients who had aneurysms involving their aortic arch and/or descending aorta (group B). Of the 20 cases of VCP, 19 involved the left vocal cord. The remaining one was a case of bilateral VCP. Further analysis then was performed to determine which of the demographic and operative variables listed in Table 1 were associated with the 3 outcomes. Among the patients who had aneurysms limited to the aortic root and/or ascending aorta (group A), an increasing duration of CPB time emerged as a significant factor associated with the risk of VCP and death (p ⫽ 0.0002) (Table 3). Within group B (ie, procedures involving the aortic arch and/or descending aorta), 2 factors (ie, the duration of CPB and descending aneurysm) emerged as statistically significant independent variables associated with VCP and death (CPB time, p ⫽ 0.002; descending aneurysm, p ⫽ 0.03; Fig 1). None of the additional variables examined approached statistical significance within either group of patients.
Table 3. Vocal Cord Paralysis and Death by the Duration of CPB Time Among Group A Patients Duration of CPB Time (min)
ⱕ200 200-240 240-300 ⬎300
Vocal Cord Paralysis (%)
Death (%)
Neither (%)
0 (0) 0 (0) 1 (1.11) 1 (1.85)
0 (0) 2 (2.78) 0 (0) 5 (9.26)
34 (100) 70 (97.2) 89 (98.8) 48 (88.8)
NOTE. p ⫽ 0.0002. Group A: aneurysms involving the aortic root and/or ascending aorta. Abbreviation: CPB, cardiopulmonary bypass.
524
DILISIO ET AL
Fig 1. Independent risk factors (ie, cardiopulmonary bypass time and descending aortic aneurysm) associated with clinical outcomes among patients within group B. Vocal cord paralysis (), death ( ), and neither ( ). Cardiopulmonary bypass time: p ⴝ 0.002, descending aortic aneurysm: p ⴝ 0.03. Group B: aneurysms involving the aortic arch and/or descending aorta.
To study the postoperative impact of VCP, the authors compared the LOS of the 3 outcome groups (ie, VCP, death, and neither) within each surgical group (Fig 2). Within group A, the median (minimum, maximum) LOS for patients who had VCP was 20.5 days (13, 28 days). The median (minimum, maximum) LOS for the patients who died was 12 days (0, 182 days). The survivors without vocal cord paralysis (VCP) had a median (minimum, maximum) LOS of only 8 days (4, 114 days). Among the group B patients, the median (minimum, maximum) LOS for the VCP group also was 20.5 days (7, 112 days). The median (minimum, maximum) LOS for the patients who died was 9.5 days (1, 101 days). The survivors without VCP had a median (minimum, maximum) LOS of 12 days (4, 182 days). The LOS differences were statistically significant (p ⫽ 0 .017) for group A patients but failed to reach statistical significance in group B patients (p ⫽ 0.059). Among the group A patients, the number (%) with a tracheostomy was 1 (50%), 1 (14.3%), and 4 (1.7%) for VCP, death, and neither, respectively (p ⫽ 0.007). The corresponding numbers for group B patients were 4 (22.2%), 4 (20.0%), and 18 (8.6%), respectively (p ⫽ 0.057, Fig 3). DISCUSSION
The authors have shown a statistically significant association between the increased length of hospital stay and tracheostomy among patients who are diagnosed with vocal cord paralysis after procedures involving their aortic root and/or ascending aorta. Although this association was not statistically significant among patients who suffer the same complication after procedures involving their aortic arch and/or descending aorta, it
Fig 2. The median postoperative LOS by clinical outcomes for groups A and B. Vocal cord paralysis (), death ( ), and neither ( ). Group A: p ⴝ 0.017; group B: p ⴝ 0.059. Group A: aneurysms involving the aortic root and/or ascending aorta. Group B: aneurysms involving the aortic arch and/or descending aorta.
clearly trended in the same direction (p ⫽ 0.059 for LOS and p ⫽ 0.057 for tracheostomy). Among all patients undergoing thoracic aortic surgery, increasing the duration of CPB time was associated with an increased risk of vocal cord paralysis. Independent of the duration of the CPB time, procedures performed on the descending thoracic aorta were also associated with an increased risk of vocal cord paralysis. This is the largest study to date looking specifically at VCP in adults after thoracic aortic surgery. The incidence of VCP was found to be the highest (7.26%) in patients who presented for the repair of aneurysms that involved their aortic arch and/or descending aorta. This percentage was similar to other studies that have reported an incidence of near 9% in procedures at this location.4,5 Other studies that focused primarily on coronary artery bypass graft procedures without the involvement of the ascending aorta reported a risk near 2%.2,6 This is similar to the group A patients despite the fact that additionally all ascending aortic repairs were included in group A. Clearly, the anatomic course of the left recurrent laryngeal nerve as it branches from the vagus nerve at the level of the transverse aortic arch and loops under it, posterior to the ligamentum arteriosum, to ascend up the thorax places it at risk during thoracic aortic surgical procedures. The fact that nearly all cases of VCP in this study involved only the left vocal cord further supports the theory that the nerve was damaged during the surgical procedure itself. However, there are other causes of VCP after surgery. One retrospective study of 31,241 patients reported that 0.08% of patients suffered from VCP simply after tracheal intubation.7 Although this retrospective study did show an increased asso-
VOCAL CORD PARALYSIS
525
Fig 3. The incidence of tracheostomy ( ) and no tracheostomy ( ) within clinical outcomes for groups A and B. Group A: p ⴝ 0.007. Group A: aneurysms involving the aortic root and/or ascending aorta. Group B: p ⴝ 0.057. Group B: aneurysms involving the aortic arch and/or descending aorta.
ciation with prolonged intubation time and increasing age, other prospective investigations have disputed this finding.8 Increased pressure from an inflated endotracheal tube cuff transmitted to the tracheal mucosa near where the external branch of the recurrent laryngeal nerve enters the trachea to innervate the vocal cords is thought to be the mechanism behind intubation-related injury. A condition sometimes mistaken for VCP is arytenoid cartilage dislocation. Rubin et al9 reported that this is an extremely rare complication of traumatic intubation and that it is distinguished from true VCP with strobovideolaryngoscopy. The present authors believe that because all 18 cases of VCP diagnosed in patients who had procedures involving the aortic arch and/or descending aorta were involving the left vocal cord that the respective nerve was injured during the surgical repair. Although damage to the recurrent laryngeal nerve during the surgical procedure is the most likely culprit, it is interesting to note that none of the dictated operative reports mentioned this complication. One investigation was able to show that if the operating surgeon had the impression that the recurrent laryngeal nerve possibly was damaged during surgery that it was in fact associated with a higher probability of vocal cord immobility.10 This may be evidence enough that instead of relying on the operative note, it would be worthwhile to simply ask the operating surgeon if they believe there was a chance that the recurrent laryngeal nerve was damaged during surgery to anticipate potential associated complications.
The finding that the risk of VCP increased with an increasing duration of CPB time in both groups is not surprising. Two other studies have shown that the risk of VCP increased with an increasing duration of both CPB time and total operative time.1,11 This may be a marker for a more complex dissection that makes injury to the nerve more likely. Alternatively, lengthy procedures and increasing the duration of CPB also may be a marker for a longer period during which patients are exposed to colder temperatures, which is another known risk factor for nerve injury. This type of nerve injury often is limited to the immediate time after rewarming and not likely a contributor to permanent damage unless hypothermia is severe and prolonged.12 However, there are case reports of injury in this manner to both the phrenic and recurrent laryngeal nerves producing diaphragmatic and VCP, respectively, lasting nearly 1 year.13 The injury was attributed to a prolonged exposure to ice/slush mix used for myocardial preservation. In the present investigation, no attempt was made to address other potential causes of nerve damage. If hypothermia was a significant source of permanent nerve injury, clinicians would expect to see an association between the incidence of VCP and the duration of an esophageal temperature time below 20°C or a duration of deep hypothermic circulatory arrest. The present authors did not find such associations in their analyses. The present results showed a strong association between the location of the procedure and the risk of VCP. Specifically, in the present study, the risk of experiencing VCP in the group of
526
DILISIO ET AL
patients with aortic arch and/or descending aortic aneurysm repairs (group B) was at least 7.3% compared with only 0.8% for the group A patients, who had aneurysms involving the aortic root and ascending aorta. Classifying the group B patients further in terms of whether the aortic arch or descending thoracic aorta was involved in the actual surgical procedure provides a deeper understanding of how the anatomic location of the pathology influences the risk of VCP. Because of the location of the recurrent laryngeal nerve and the aortic arch, it was not unexpected to find that the risk associated with procedures involving descending aortic aneurysms was significantly greater than with the ascending aorta and/or aortic arch (p ⬍ 0.03). Other reports also have shown that the incidence of postoperative left-sided vocal cord paralysis is greater in patients who undergo repair limited to the descending aorta or distal arch aneurysms.1,11 Ohta and Mori,14 in a series examining 69 patients who underwent surgery on their descending thoracic aorta, found that chronic dilation at the left subclavian artery and surgical anastomosis occurring proximal to the left subclavian artery were independent risk factors for postoperative VCP (odds ratios ⫽ 8.67 and 17.7, respectively). It is quite possible that within the 110 descending aortic aneurysms in the present series the surgeon performed the anastomosis proximal to the left subclavian artery because of the immediacy of the aneurysmal formation following the distal edge of this artery. The cumulative weight of the previously mentioned evidence showed that VCP in the postoperative period was the highest when the pathology and surgical site were located at or near the origin of the left subclavian artery. The association between VCP and a greater length of hospital stay was similar to what other investigators have found.1 This association reached statistical significance among patients who had aneurysms involving the aortic root and/or ascending aorta and approached statistical significance in the group of patients who had procedures that involved the aortic arch and/or descending aorta. Although the additional cost incurred per hospital day is difficult to estimate, clearly a longer LOS further strains an already overburdened health care system.15 The dramatic association between VCP and tracheostomy is cause for concern. The present study showed a 25% incidence of tracheostomy among all patients who had VCP compared with the 5% incidence in the survivors without VCP. Ohta et al11 reported a 22.5% incidence of tracheostomy or reintubation in their cohort of 40 patients who suffered from VCP. Although the present study did not examine additional complications such as swallowing difficulty, decreased cough reflex, and aspiration pneumonia that potentially are associated with VCP leading to tracheostomy, there are multiple studies in both the
pediatric and adult cardiothoracic literature confirming this association.1,16,17 The aryepiglottic muscle, which receives innervation from the recurrent laryngeal nerve, constricts during swallowing, protecting the entrance to the larynx.18 Damage to the recurrent laryngeal nerve exposes the larynx during swallowing, increasing the risk of aspiration. There is evidence that the early treatment of known VCP by injection laryngoplasty reduces the risk of developing aspiration pneumonia.19 A majority of the patients who suffered VCP in the present study did have injection laryngoplasty performed; however, it is unknown if this procedure averted any cases of aspiration pneumonia because these data were unavailable. Aspiration also has been shown to be a risk factor for tracheostomy in a large prospective cohort study of 552 intensive care patients.20 Tracheostomy has been associated with declines in physical and mental functioning from the patient perspective and is viewed widely as a setback in the recovery period.21 This study was limited by the fact it was a retrospective study. Also, the vocal cord status of the 27 patients who died was not confirmed. These data were very difficult to discover because many patients died before they were extubated, and could not be examined for vocal cord mobility. The incidence of VCP in the present cohort was likely underestimated because asymptomatic patients did not undergo evaluation by an otolaryngologist. One study that looked prospectively at the incidence of VCP in children undergoing cardiothoracic surgery showed that 14% of asymptomatic patients suffered from VCP.22 Additionally, some perioperative variables, such as aneurysm size and surgical technique, which may have an association with VCP, were not examined. Despite these limitations, the authors believe these results are significant in light of the limited studies available regarding this topic. Once thought of as a relatively minor complication in comparison with other more serious outcomes such as renal failure, paralysis, or sepsis in this surgical population, VCP may portend another important prognosis regarding tracheostomy and increased LOS. One study of patients undergoing esophagectomy found unilateral VCP in 34% of patients. However, this outcome was not associated with a longer ICU stay or an increase in pulmonary complications.23 The present authors have shown that the same injury to the recurrent laryngeal nerve in thoracic aortic surgical patients does have an association with greater morbidity. ACKNOWLEDGMENT The authors would like to thank John Essad for his help with table and graph editing.
REFERENCES 1. Itagaki T, Kikura M, Sato S: Incidence and risk factors of postoperative vocal cord paralysis in 987 patients after cardiovascular surgery. Ann Thorac Surg 83:2147-2152, 2007 2. Shafei H, El-Kholy A, Azmy S, et al: Vocal cord dysfunction after cardiac surgery: an overlooked complication. Eur J Cardiothorac Surg 11:564-566, 1997 3. Ishimoto S, Ito K, Toyama M, et al: Vocal cord paralysis after surgery for thoracic aortic aneurysm. Chest 121:1911-1915, 2002
4. Ishii K, Adachi H, Tsubaki K, et al: Evaluation of recurrent nerve paralysis due to thoracic aortic aneurysm and aneurysm repair. Laryngoscope 114:2176-2181, 2004 5. Bhaskar J, Foo J, Sharma A. Clamp-and-sew technique for traumatic injuries of the aorta: 20-year experience. Asian Cardiovasc Thorac Ann 18:161-165, 2010 6. Hamdan A, Moukarbel R, Farhat F, et al: Vocal cord paralysis after open-heart surgery. Eur J Cardiothorac Surg 21:671-674, 2002
VOCAL CORD PARALYSIS
7. Kikura M, Suzuki K, Itagaki T, et al: Age and comorbidity as risk factors for vocal cord paralysis associated with tracheal intubation. Br J Anaesth 98:524-530, 2007 8. Colton J, Noordzij J, Murgia B, et al: Laryngeal injury from prolonged intubation: a prospective analysis of contributing factors. Laryngoscope 121:596-600, 2011 9. Rubin A, Hawkshaw M, Moyer C, et al: Arytenoid cartilage dislocation: A 20-year experience. J Voice 19:687-701, 1997 10. Carpes L, Kozak F, Leglanc J, et al: Assessment of vocal fold mobility before and after cardiothoracic surgery in children. Arch Otolaryngol Head Neck Surg 137:571-575, 2011 11. Ohta N, Kuratani T, Hagihira S, et al: Vocal cord paralysis after aortic arch surgery: predictors and clinical outcome. J Vasc Surg 43:721-728, 2006 12. Stecker M, Baylor K: Peripheral nerve at extreme low temperatures 1: effects of temperature on the action potential. Crybiology 59:1-11, 2009 13. Tewari P, Aggarwal S: Combined left-sided recurrent laryngeal and phrenic nerve palsy after coronary artery operation. Ann Thorac Surg 61:1721-1723, 1996 14. Ohta N, Mori T: Vocal cord paralysis after surgery to the descending thoracic aorta via left posterolateral thoracotomy. Ann Vasc Surg 21:761-766, 2007 15. Seidel J, Whiting P, Edbrooke D: The costs of intensive care. Continuing Educ Anaesth Crit Care Pain 6:160-163, 2006
527
16. Sachdeva R, Hussain E, Moss MM, et al: Vocal cord dysfunction and feeding difficulties after pediatric cardiovascular surgery. J Pediatr 151:312-315, 315.e1-2, 2007 17. Benjamin J, Smith B, Cotton M, et al: Long-term morbidities associated with vocal cord paralysis after surgical closure of a patent ductus arteriosus in extremely low birth weight infants. J Perinatol 30:408-413, 2010 18. Thach B: Recurrent laryngeal nerve injury during patent ductus ligation: Can this common complication be reduced? J Perinatol 30: 371-372, 2010 19. Graboyes E, Bradley J, Meyers B, et al: Efficacy and safety of acute injection laryngoplasty for vocal cord paralysis following thoracic surgery. Laryngoscope 212:2406-2410, 2011 20. Kollef M, Ahrens T, Shannon W: Clinical predictors and outcomes for patients requiring tracheostomy in the intensive care unit. Crit Care Med 27:1714-1720, 1999 21. Hashmi N, Ransom E, Nardone H, et al: Quality of life and self-image in patients undergoing tracheostomy. Laryngoscope 120: S196, 2010 22. Smith M, King J, Elsherif A, et al: Should all newborns who undergo patent ductus arteriosus ligation be examined for vocal fold mobility? Laryngoscope 119:1606-1609, 2009 23. Johnson P, Kanegoanker G, Bates T: Indirect laryngoscopic evaluation of vocal cord function in patients undergoing transhiatal esophagectomy. J Am Coll Surg 178:605-608, 1994