Resolution of Laryngeal Injury Following Translaryngeal Intubation 1,2
GENE L. COLICE
Introduction SUMMARY Translaryngeal Intubation (TLI) causes mucosal ulcerations of the vocal conls and
Tracheal tubes placed through either the posterior laryngeal commissure. Usually these ulcers heal by primary reeplthellallzatlon, but occanose or mouth and translaryngeally are sionally laryngeal granulomas or strictures develop at these ulcer sites. The Incidence of granuloma forced against the posterior larynx by the and stricture formation and the variables Influencing abnormal laryngeal healing following TLI are tongue and the curvature of the cervical not well understood. A group of 54 patients who experienced prolonged TLiwere foll~d prospecspine. This pressure, along with tube tively to determine the resolution rate of laryngeal Injury. Direct flberoptlc laryngoscopy was performed at either extubation or tracheostomy and repeated every 2 wk until the larynx returned to movement caused by swallowing, respinormal or a persistent laryngeal abnormality was Identified. Laryngeal symptoms were assessed ration, attempted phonation, and neck at these same time points. In 5 patients (9%) the appearance of the larynx was normal at extubation, motion, mechanically abrades the larynx. and In 42 patients (78%) laryngeal healing occurred by primary reeplthellallzatlon within 8 wk. Four At extubation mucosal ulcerations and patients (7%) developed laryngeal granulomas, which required surgical removelln all but one case. edema are commonly observed along the No patients In this series developed laryngeal strictures. Three patients (6%) died before complete posterior aspects of both the true vocal follOW-Up. Laryngeal symptoms, pertlcularly hoarseness, resolved as the larynx healed. Performance cords and the posterior laryngeal comof traCheostomy, age, TLI for more than 10 days, and severe laryngeal InjUry at extubation did not missure (1,2). Healing of these mucosal influence the median tl'me to resolution of laryngeal abnormalities. Abnormal laryngeal healing folulcers may occur through primary reepilowing TLI is uncommon but Is not exacerbated by prolonged TLI (more than 10days), severe larynthelialization or secondarily via granugeal Injury at extubation, or performance of a tracheostomy. AM REV RESPIR DIS 1992; 145:361-364 loma formation (2). Occasionally, scarring and wound contraction occur (3, 4). Why some patients heal by granuloma formation or scarring, rather than primary reepithelialization, is not underfice area during inspiration by more than 50070 Methods stood. Some reports have suggested that or laryngospasm). Every2 wk followingstudy Patients prolonged translaryngeal intubation entry the patients underwent repeat laryn(TLI) (3, 4), performance of secondary The methods used in this study have already goscopy and laryngeal damage was again been described in detail (1). All patients adtracheostomy (1), and severelaryngeal in- mitted to the medical, surgical, and respira- graded according to these criteria. At these jury at extubation (3, 4) predispose to ab- tory intensive care units (ICU) at the Tampa examinations laryngeal symptoms, for examnormal laryngeal healing. Determining VAHospital between September 15, 1982and ple hoarseness, difficulty swallowing, throat whether these variables play such a role December 31, 1984 who had TLI for more pain, or stridor, were also assessed. Followup continued until the laryngeal examination is clinically important, because laryngeal than 4 consecutive days, no history of prior was normal, the patient died, or abnormal laryngeal pathology, and a life expectancy of healing by wayof either granuloma or scar laryngeal healing was recognized. formation may lead to chronic hoarse- more than 2 wk were identified by the author (GLC). Within 24 h of removal of the transness, airway narrowing, or even respiraPatient Care laryngeal tube, either byextubation or replacetory failure. Medical care throughout the course of the paWe prospectively evaluated in a two- ment by a tracheostomy, an informed con- tient's illness was provided by house officers sent approved by the institutional review phase study laryngeal complications in board was signed and the patient was entered and attending staff from the University of South Florida College of Medicine. Transa large group of patients who had TLI into the study. laryngeal intubations were primarily perfor more than 4 days. In the first phase Immediately upon entry into the study (and of this study weexamined laryngeal dam- within 24 h of extubation or tracheostomy) age at extubation and evaluated how such direct laryngoscopy was performed (by GLC) damage might be related to clinically rec- using a flexible fiberoptic bronchoscope. (Received in original form October 15, 1990 and ognized laryngeal complications (1). In Laryngeal damage was graded by visual ex- in revised form April 17, 1991) the second phase, the results of which are amination as none (no damage), mild (presFrom the Department of Internal Medicine, presented here, we studied laryngeal heal- ence of erythema and mucosal ulcerations, Dartmouth Medical School, Hanover, New Hampbut no reduction in laryngeal orifice sizeduring. This allowed us to examine whether ing inspiration), moderate (presence of ery- shire, and the White River Junction Veterans Adduration of TLI, performance of second- thema, mucosal ulcerations, and mucosal ministration Hospital, White River Junction, ary tracheostomy, or severity of laryn- swelling reducing laryngeal orifice area dur- Vermont. 2 Correspondence and requests for reprints geal injury at extubation influenced the ing inspiration by less than half), or severe should be addressed to Gene L. Colice,M.D., Chief, resolution rate and outcome of laryngeal (presence of erythema, mucosal ulcerations, Pulmonary, VA Hospital, White River Junction, healing. and mucosal swelling reducing laryngeal oriVT 05001. 1
361
362
GENE L. COLICE
TABLE 1 PATIENTS NOT DEMONSTRATING COMPLETE LARYNGEAL HEALING' Duration TLI
Grade Initial Laryngeal Injury
Reason
Week
2 1 2 1 3 2
Laryngeal granuloma Laryngeal granuloma Died. aspiration pneumonia Laryngeal granuloma Died, aspiration Died. aspiration pneumonia
12 12 4 20 2 2
8 10 21
8 6 5
• One patient developed a laryngeal granuloma and was transiently lost to follow-up at Week 16 but on reassessment at Week 24 was found to have had spontaneous resolution of the granuloma.
formed by house officers, who had widely varying skills in this technique. Respiratory therapists provided the same respiratory care techniques in all three intensive care units. Tracheal tubes placed translaryngeally were taped securely in position close to the mouth or nose. All tracheal tubes were 2-79 approved and single use. Two-way swivel adaptors and elastic latex connectors were interposed between the tracheal tube and attached tubing. Attached tubing was kept to a minimum and suspended by support arms. Suctioning of tracheal tubes was performed as clinically indicated with a standard sterile glove technique and disposable plastic catheters. Tracheal tube balloon cuffs were inflated to the minimal leak point and were kept inflated 24 h a day. Balloon cuff pressures were monitored every 8 h, and tracheal tubes with cuffs that required more than 35 mm Hg pressure for proper sealing were replaced. Tracheostomy wound dressings were changed as often as necessary to keep the stoma dry. Tracheostomy was performed when deemed clinically indicated by the appropriate house officers and staff physicians. Laryngoscopy results obtained as part of this study were not used as a basis for clinical decisions regarding extubation or tracheostomy.
Statistical Analysis The median time to resolution of laryngeal pathology was calculated for patients categorized in the following ways: tracheostomy (performed or not), severity of laryngeal injury seen at initial laryngoscopy (graded as mild, moderate, or severe), duration of TLI (10 days or more versus 4 to 9 days), and age (less than 55, 55 to 64, 65 to 74, and greater than 75 yr), Differences among categories with respect to median resolution time were tested individually by the logarithmic rank test (5) and, adjusted for other covariates, by the Cox proportional hazards model (6). Differences were considered significant when p < 0.05.
Results
Eight-six consecutive patients meeting the entry criteria were screened: Twopatients werenot entered because of incomplete records. One patient declined participation, and one patient was not en-
tered because of high-grade ventricular ectopy. Eighty-two patients were entered into the original study (1), but 28 patients either died within 2 wk of extubation or required reintubation for medical problems unrelated to their larynx. Fifty-four patients were followed prospectively. They were all men, had a mean age of 63.1 ± 10.9(SD) yr, and experienced TLI for a mean 9.3 ± 5.1 (SD) days. None was lost to follow-up, and 94070 of all laryngeal examinations were completed as scheduled. The median time to resolution of laryngeal pathology was 4 wk. Actuarial estimates of the time to resolution of laryngeal symptoms closely approximated those for resolution of laryngeal pathology, with a median resolution time of 4 wk. Five patients (9%) had a normal laryngeal examination at extubation. Forty-two patients (78%) had laryngeal injury at extubation which healed by primary reepithelialization within 8 wk. Four patients (7%) healed secondarily through granuloma formation. In one of these patients the granulomas eventually disappeared spontaneously 24 wk after extubation. The other three patients (6%) had chronic hoarseness from laryngeal granulomas and required surgical intervention. No patient in this series developed laryngeal scarring. Three patients (6%) died before complete follow-up. The six cases who did not demonstrate complete healing are summarized in table l. Figure 1 indicates actuarial estimates of the time to resolution of laryngeal injury categorized by whether the patient had a tracheostomy (see table 2 for patient characteristics), severity of initial laryngeal inury (grade mild, moderate, or severe), and duration of TLI (10 days or more versus 4 to 9 days). The median times for resolution of laryngeal pathologyare similar for all categories. The median resolution time (data not shown) did not differ among the various age groups. None of the patients who underwent tra-
cheostomy or who had severe initial laryngeal injury healed secondarily by granuloma formation. Only one patient with TLI for 10 days or more developed laryngeal granulomas. Discussion
Wepreviously presented data from a larger patient group on the variables associated with severity of laryngeal injury at extubation and clinically recognizable laryngeal complications of such injury (1). In this study we prospectively examined whether certain variables influenced the resolution of laryngeal injury following extubation. We found that the vast majority of these patients (47 of 51, or 92070) had a normal laryngeal examination by 8 wk following extubation from TLI and only four patients (7%) healed their laryngeal injury secondarilythrough granuloma formation. Resolution of laryngeal symptoms, particularly hoarseness, closely matched laryngeal healing. Duration of TLI, performance of a tracheostomy, and severity of initiallaryngeal injury did not predispose to laryngeal scarring or granuloma formation in this study. Most previous studies evaluating the resolution of laryngeal injury following TLI are of limited value because the patients had short-duration TLI, follow-up was incomplete, or the effect of tracheostomy was not adequately assessed. Careful evaluation of laryngeal healing following TLI for general anesthesia has shown a less than 1% incidence of laryngeal scarring (7, 8). The duration of TLI in these patients was quite short, however, probably only hours. These findings may, therefore, not be applicable to present-day intensive care unit patients who require TLI for up to several weeks. Others (9-13) have assessed laryngeal complications in critically ill patients requiring TLI and tracheostomy for longer periods of time and have also reported a small overall incidence of either laryngeal granulomas or stenosis. Unfortunately, direct visual examination of the larynx was not performed in a systematic manner in these studies and follow-up of these patients was inconsistent. Consequently, a true incidence of abnormal laryngeal healing cannot be determined from these studies. Kastonas and colleagues (14), Burns and coworkers (15), and Lindholm (2) completed careful studies in which seriallaryngeal examinations following extubation allowed an accurate assessment of laryngeal healing. Kastonas and col-
RESOWTION OF lARYNGEAL INJURY AFTER
ru
1.0
363 a
TABLE 2 CHARACTERISTICS OF PATIENTS UNDERGOING TRACHEOSTOMY·
09 08
TLI Alone
TLI Plus Tracheostomy
41 63.4 ± 11.6 8.0 ± 4.4
62.2 ± 8.8 13.6 ± 4.8t
1.4 ± 0.8
2.4 ± 0.9t
07 06
Number Age Duration TLI Grade initial laryngeal injury
05 0.4 I
- - --~
J.'
0"
13
• Values are means ± SD.
t p < 0.01 from TLI alone.
oJ 12
0
16
20
24
Weeks
'0
Fig. 1. Resolution rates of laryngeal injury are categorized by the following variables. (a) Patients are divided into groups based on whether a secondary tracheostomy was performed (solidline) or not (dashed line). (b) Categories are based on severity of laryngeal injury at extubation: mild (solid line); moderate (short-dashed line). severe (long-dashed line). (c) Groupings are determined by duration of TLI between 4 and 9 days (solid line) or 10 or more days (dashed
b
0.9 0.8
c------------------------ ------ -------- ----
0.7
~
. ----4 I
__ J ~
0.5
go
0.4
Q:
03 0.2
line).
01
12
16
20
24
16
20
24
Weeks
1.0
C
0.9 08 07 ~
~ 0.6
e
005
;:
8.
£.
0.4 j----
0.3 0.2 0.1
12
Weeks
leagues (14)found 2 of 19 patients (11 070) to have vocal cord granulomas 16wk after extubation. Burns and coworkers (15) detected no laryngeal granulomas in 117 patients by 12wk after extubation. Lindholm (2) carefully studied 69 adults by laryngoscopy every 15days following extubation. Of these patients, 43 healed the larynx by primary reepithelialization within 4 wk. The other 26 patients developed initial granulomas at the site of laryngeal ulcers, but in 23 patients these granulomas spontaneously resolved by a median time of 9 wk. Two patients re-
quired surgical removal of granulomas, and one patient had surgical correction of a posterior commissure fibrous band. These studies agreed that laryngeal healing is usually completed by 9 to 16 wk after extubation with a low rate of scarring or persistent granulomas. A limiting factor in all these studies, however, was failure to adequately define the influence of tracheostomy on resolution of laryngeal injury. Both Burns and coworkers (15) and Lindholm (2) studied patients who underwent tracheostomy following TLI, but did not separately report the rate
of abnormal laryngeal healing in these patients. Kastonas and colleagues(14) did not include patients receiving tracheostomy in their study. Whited studied laryngeal healing in 200 patients who had TLI for 2 to 24 days (3,4). The majority of these patients had a normal larynx by 3 wk after extubation, with healing seemingly faster with shorter duration TLI. Twelve (6%) of these patients developed laryngeal scarring by 16 wk following extubation. In 11 patients scarring was observed in the interarytenoid plane and inferiorly along the internal face of the cricoid. The other patient developed fibrous bands between the vocal cords. Although all 12 of these patients had undergone tracheostomy in addition to TLI, Whited believed that tracheostomy was not an important variable promoting abnormal laryngeal healing. Instead, he speculated that the severity of laryngeal injury at extubation was the critical factor influencing the development of laryngeal scarring. This study has avoided the limitations of previous studies. Patients experienced prolonged TLI before extubation or tracheostomy, reflecting general practice in present-day intensive care units. Followup was complete for all patients and direct visual examination of the larynx was performed in a prospective, systematic manner. This allows the true rate of laryngeal healing to be expressed. Patients undergoing tracheostomy were included to allow evaluation of this procedure's effect on resolution of laryngeal injury. Unlike Whited, we found no cases of laryngeal scarring and a small rate (3 of 51, or 6%) of persistent laryngeal granulomas. Healing was usually complete by 8 to 12wk. These results are similar to those described by others (2, 14, 15). These observations have important clinical implications. Because severity of laryngeal injury at extubation did not influence either the rate or manner of
364
laryngeal healing, laryngoscopy findings at extubation should not be used to determine whether tracheostomy is needed to minimize the development of future laryngeal pathology. Laryngoscopy should be considered, however, in patients who have hoarseness persisting beyond the median time to resolution of laryngeal pathology, which was 4 wk in this study. Surgically correctable problems, such as laryngeal granulomas, may be found in these cases. The finding that duration of TLI did not influence laryngeal healing, in conjunction with our previous observation that duration of TLI does not appear to affect severity of laryngeal injury at extubation, disputes the commonly held viewpoint that secondary tracheostomy is advisable in patients requiring prolonged TLI for mechanical ventilation to protect the larynx, Unfortunately, this study does not provide insights into what variables promote abnormal laryngeal healing. Besides the factors detailed in this study, no other clinically obvious common theme, such as aspiration, poor nutrition, or type of underlying disease, was found in the patients who developed laryngeal granulomas. Presumably, individual variability in wound healing, a poorly defined but
GENE L. COLICE
well-recognized concept, determined whether laryngeal injury healed by primary reepithelialization. However, this study reassures the clinician that abnormal laryngeal healing, by either scarring or granuloma formation, occurs much less often and with less morbidity than recently reported by others (3, 4). In conclusion, laryngeal injury at extubation is common, but healing by primary reepithelialization usually occurred within 8 wk. Few patients developed laryngeal granulomas, and no cases of laryngeal scarring wereobserved. Tracheostomy, TLI lasting more than 10 days, and severelaryngeal injury at extubation did not influence laryngeal healing. Acknowledgment The author appreciates the advice on statistical evaluation of the data provided by Therese Stukel, Ph.D. and Bradley Dain, M.S., Biostatistical Shared Service; Darmouth Medical School.
References 1. Colice GL, Stukel TA, Dain B. Laryngeal complications of prolonged intubation. Chest 1989; 96:877-84.
2. Lindholm CEo Prolonged endotracheal intubation. Acta Anaesthesiol Scand 1969; 33:1-131. 3. Whited RE. Posterior commissurestenosispost
long-term intubation. Laryngoscope 1983; 93: 1314-8.
4. Whited RE. A prospective study of laryngotracheal sequelae in longer-term intubation. Laryngoscope 1984; 94:367-77. 5. Kalbfleisch ID, Prentice RL. The statistical analysisof failuretime data. NewYork:John Wiley and Sons, 1980. 6. Cox DR. Regression models and life tables. 1 R Stat SOC IB] 1972; 34:187-220. 7. Kambic V, Radsel Z. Intubation lesion of the larynx. Br 1 Anaesth 1978; 50:587-90. 8. Arner 0, Diamant M. Respiratory-tract lesions following intratracheal anaesthesia. Acta Chir Scand 1951; 101:75-84. 9. Aass AS. Complications to tracheostomy and long-term intubation: a follow-up study. Acta Anaesthesiol Scand 1975; 19:127-33. 10. Stauffer Jl., Olson DE, Petty TL. Complications and consequences of endotracheal intubation and tracheostomy. Am 1 Med 1981; 70:65-76. 11. Tonkin IP, Harrison GA. The effect on the larynx of prolonged endotracheal intubation. Med 1 Aust 1%6; 2:581-7. 12. Dixon TC, Sando MIW, Bolton 1M, Gilligan lE. A report of 342 casesof prolonged endotracheal intubation. Med 1 Aust 1968; 2:529-33. 13. Dunham CM, LaMonica C. Prolonged intubation in the trauma patient. 1 Trauma 1984; 24:120-4. 14. Kastanos N, Miro RE, PerezAM, etal. Laryn-
gotracheal injury due to endotracheal intubation. Crit Care Med 1983; 11:362-7. 15. Burns HP, Dayal VS, Scott A, et al. Laryngotracheal trauma: observations on its pathologenesis and its prevention following prolonged orotracheal intubation in the adult. Laryngoscope 1979; 89:1316-24.
GENE L. COLICE
Introduction SUMMARY Translaryngeal Intubation (TLI) causes mucosal ulcerations of the vocal conls and
Tracheal tubes placed through either the posterior laryngeal commissure. Usually these ulcers heal by primary reeplthellallzatlon, but occanose or mouth and translaryngeally are sionally laryngeal granulomas or strictures develop at these ulcer sites. The Incidence of granuloma forced against the posterior larynx by the and stricture formation and the variables Influencing abnormal laryngeal healing following TLI are tongue and the curvature of the cervical not well understood. A group of 54 patients who experienced prolonged TLiwere foll~d prospecspine. This pressure, along with tube tively to determine the resolution rate of laryngeal Injury. Direct flberoptlc laryngoscopy was performed at either extubation or tracheostomy and repeated every 2 wk until the larynx returned to movement caused by swallowing, respinormal or a persistent laryngeal abnormality was Identified. Laryngeal symptoms were assessed ration, attempted phonation, and neck at these same time points. In 5 patients (9%) the appearance of the larynx was normal at extubation, motion, mechanically abrades the larynx. and In 42 patients (78%) laryngeal healing occurred by primary reeplthellallzatlon within 8 wk. Four At extubation mucosal ulcerations and patients (7%) developed laryngeal granulomas, which required surgical removelln all but one case. edema are commonly observed along the No patients In this series developed laryngeal strictures. Three patients (6%) died before complete posterior aspects of both the true vocal follOW-Up. Laryngeal symptoms, pertlcularly hoarseness, resolved as the larynx healed. Performance cords and the posterior laryngeal comof traCheostomy, age, TLI for more than 10 days, and severe laryngeal InjUry at extubation did not missure (1,2). Healing of these mucosal influence the median tl'me to resolution of laryngeal abnormalities. Abnormal laryngeal healing folulcers may occur through primary reepilowing TLI is uncommon but Is not exacerbated by prolonged TLI (more than 10days), severe larynthelialization or secondarily via granugeal Injury at extubation, or performance of a tracheostomy. AM REV RESPIR DIS 1992; 145:361-364 loma formation (2). Occasionally, scarring and wound contraction occur (3, 4). Why some patients heal by granuloma formation or scarring, rather than primary reepithelialization, is not underfice area during inspiration by more than 50070 Methods stood. Some reports have suggested that or laryngospasm). Every2 wk followingstudy Patients prolonged translaryngeal intubation entry the patients underwent repeat laryn(TLI) (3, 4), performance of secondary The methods used in this study have already goscopy and laryngeal damage was again been described in detail (1). All patients adtracheostomy (1), and severelaryngeal in- mitted to the medical, surgical, and respira- graded according to these criteria. At these jury at extubation (3, 4) predispose to ab- tory intensive care units (ICU) at the Tampa examinations laryngeal symptoms, for examnormal laryngeal healing. Determining VAHospital between September 15, 1982and ple hoarseness, difficulty swallowing, throat whether these variables play such a role December 31, 1984 who had TLI for more pain, or stridor, were also assessed. Followup continued until the laryngeal examination is clinically important, because laryngeal than 4 consecutive days, no history of prior was normal, the patient died, or abnormal laryngeal pathology, and a life expectancy of healing by wayof either granuloma or scar laryngeal healing was recognized. formation may lead to chronic hoarse- more than 2 wk were identified by the author (GLC). Within 24 h of removal of the transness, airway narrowing, or even respiraPatient Care laryngeal tube, either byextubation or replacetory failure. Medical care throughout the course of the paWe prospectively evaluated in a two- ment by a tracheostomy, an informed con- tient's illness was provided by house officers sent approved by the institutional review phase study laryngeal complications in board was signed and the patient was entered and attending staff from the University of South Florida College of Medicine. Transa large group of patients who had TLI into the study. laryngeal intubations were primarily perfor more than 4 days. In the first phase Immediately upon entry into the study (and of this study weexamined laryngeal dam- within 24 h of extubation or tracheostomy) age at extubation and evaluated how such direct laryngoscopy was performed (by GLC) damage might be related to clinically rec- using a flexible fiberoptic bronchoscope. (Received in original form October 15, 1990 and ognized laryngeal complications (1). In Laryngeal damage was graded by visual ex- in revised form April 17, 1991) the second phase, the results of which are amination as none (no damage), mild (presFrom the Department of Internal Medicine, presented here, we studied laryngeal heal- ence of erythema and mucosal ulcerations, Dartmouth Medical School, Hanover, New Hampbut no reduction in laryngeal orifice sizeduring. This allowed us to examine whether ing inspiration), moderate (presence of ery- shire, and the White River Junction Veterans Adduration of TLI, performance of second- thema, mucosal ulcerations, and mucosal ministration Hospital, White River Junction, ary tracheostomy, or severity of laryn- swelling reducing laryngeal orifice area dur- Vermont. 2 Correspondence and requests for reprints geal injury at extubation influenced the ing inspiration by less than half), or severe should be addressed to Gene L. Colice,M.D., Chief, resolution rate and outcome of laryngeal (presence of erythema, mucosal ulcerations, Pulmonary, VA Hospital, White River Junction, healing. and mucosal swelling reducing laryngeal oriVT 05001. 1
361
362
GENE L. COLICE
TABLE 1 PATIENTS NOT DEMONSTRATING COMPLETE LARYNGEAL HEALING' Duration TLI
Grade Initial Laryngeal Injury
Reason
Week
2 1 2 1 3 2
Laryngeal granuloma Laryngeal granuloma Died. aspiration pneumonia Laryngeal granuloma Died, aspiration Died. aspiration pneumonia
12 12 4 20 2 2
8 10 21
8 6 5
• One patient developed a laryngeal granuloma and was transiently lost to follow-up at Week 16 but on reassessment at Week 24 was found to have had spontaneous resolution of the granuloma.
formed by house officers, who had widely varying skills in this technique. Respiratory therapists provided the same respiratory care techniques in all three intensive care units. Tracheal tubes placed translaryngeally were taped securely in position close to the mouth or nose. All tracheal tubes were 2-79 approved and single use. Two-way swivel adaptors and elastic latex connectors were interposed between the tracheal tube and attached tubing. Attached tubing was kept to a minimum and suspended by support arms. Suctioning of tracheal tubes was performed as clinically indicated with a standard sterile glove technique and disposable plastic catheters. Tracheal tube balloon cuffs were inflated to the minimal leak point and were kept inflated 24 h a day. Balloon cuff pressures were monitored every 8 h, and tracheal tubes with cuffs that required more than 35 mm Hg pressure for proper sealing were replaced. Tracheostomy wound dressings were changed as often as necessary to keep the stoma dry. Tracheostomy was performed when deemed clinically indicated by the appropriate house officers and staff physicians. Laryngoscopy results obtained as part of this study were not used as a basis for clinical decisions regarding extubation or tracheostomy.
Statistical Analysis The median time to resolution of laryngeal pathology was calculated for patients categorized in the following ways: tracheostomy (performed or not), severity of laryngeal injury seen at initial laryngoscopy (graded as mild, moderate, or severe), duration of TLI (10 days or more versus 4 to 9 days), and age (less than 55, 55 to 64, 65 to 74, and greater than 75 yr), Differences among categories with respect to median resolution time were tested individually by the logarithmic rank test (5) and, adjusted for other covariates, by the Cox proportional hazards model (6). Differences were considered significant when p < 0.05.
Results
Eight-six consecutive patients meeting the entry criteria were screened: Twopatients werenot entered because of incomplete records. One patient declined participation, and one patient was not en-
tered because of high-grade ventricular ectopy. Eighty-two patients were entered into the original study (1), but 28 patients either died within 2 wk of extubation or required reintubation for medical problems unrelated to their larynx. Fifty-four patients were followed prospectively. They were all men, had a mean age of 63.1 ± 10.9(SD) yr, and experienced TLI for a mean 9.3 ± 5.1 (SD) days. None was lost to follow-up, and 94070 of all laryngeal examinations were completed as scheduled. The median time to resolution of laryngeal pathology was 4 wk. Actuarial estimates of the time to resolution of laryngeal symptoms closely approximated those for resolution of laryngeal pathology, with a median resolution time of 4 wk. Five patients (9%) had a normal laryngeal examination at extubation. Forty-two patients (78%) had laryngeal injury at extubation which healed by primary reepithelialization within 8 wk. Four patients (7%) healed secondarily through granuloma formation. In one of these patients the granulomas eventually disappeared spontaneously 24 wk after extubation. The other three patients (6%) had chronic hoarseness from laryngeal granulomas and required surgical intervention. No patient in this series developed laryngeal scarring. Three patients (6%) died before complete follow-up. The six cases who did not demonstrate complete healing are summarized in table l. Figure 1 indicates actuarial estimates of the time to resolution of laryngeal injury categorized by whether the patient had a tracheostomy (see table 2 for patient characteristics), severity of initial laryngeal inury (grade mild, moderate, or severe), and duration of TLI (10 days or more versus 4 to 9 days). The median times for resolution of laryngeal pathologyare similar for all categories. The median resolution time (data not shown) did not differ among the various age groups. None of the patients who underwent tra-
cheostomy or who had severe initial laryngeal injury healed secondarily by granuloma formation. Only one patient with TLI for 10 days or more developed laryngeal granulomas. Discussion
Wepreviously presented data from a larger patient group on the variables associated with severity of laryngeal injury at extubation and clinically recognizable laryngeal complications of such injury (1). In this study we prospectively examined whether certain variables influenced the resolution of laryngeal injury following extubation. We found that the vast majority of these patients (47 of 51, or 92070) had a normal laryngeal examination by 8 wk following extubation from TLI and only four patients (7%) healed their laryngeal injury secondarilythrough granuloma formation. Resolution of laryngeal symptoms, particularly hoarseness, closely matched laryngeal healing. Duration of TLI, performance of a tracheostomy, and severity of initiallaryngeal injury did not predispose to laryngeal scarring or granuloma formation in this study. Most previous studies evaluating the resolution of laryngeal injury following TLI are of limited value because the patients had short-duration TLI, follow-up was incomplete, or the effect of tracheostomy was not adequately assessed. Careful evaluation of laryngeal healing following TLI for general anesthesia has shown a less than 1% incidence of laryngeal scarring (7, 8). The duration of TLI in these patients was quite short, however, probably only hours. These findings may, therefore, not be applicable to present-day intensive care unit patients who require TLI for up to several weeks. Others (9-13) have assessed laryngeal complications in critically ill patients requiring TLI and tracheostomy for longer periods of time and have also reported a small overall incidence of either laryngeal granulomas or stenosis. Unfortunately, direct visual examination of the larynx was not performed in a systematic manner in these studies and follow-up of these patients was inconsistent. Consequently, a true incidence of abnormal laryngeal healing cannot be determined from these studies. Kastonas and colleagues (14), Burns and coworkers (15), and Lindholm (2) completed careful studies in which seriallaryngeal examinations following extubation allowed an accurate assessment of laryngeal healing. Kastonas and col-
RESOWTION OF lARYNGEAL INJURY AFTER
ru
1.0
363 a
TABLE 2 CHARACTERISTICS OF PATIENTS UNDERGOING TRACHEOSTOMY·
09 08
TLI Alone
TLI Plus Tracheostomy
41 63.4 ± 11.6 8.0 ± 4.4
62.2 ± 8.8 13.6 ± 4.8t
1.4 ± 0.8
2.4 ± 0.9t
07 06
Number Age Duration TLI Grade initial laryngeal injury
05 0.4 I
- - --~
J.'
0"
13
• Values are means ± SD.
t p < 0.01 from TLI alone.
oJ 12
0
16
20
24
Weeks
'0
Fig. 1. Resolution rates of laryngeal injury are categorized by the following variables. (a) Patients are divided into groups based on whether a secondary tracheostomy was performed (solidline) or not (dashed line). (b) Categories are based on severity of laryngeal injury at extubation: mild (solid line); moderate (short-dashed line). severe (long-dashed line). (c) Groupings are determined by duration of TLI between 4 and 9 days (solid line) or 10 or more days (dashed
b
0.9 0.8
c------------------------ ------ -------- ----
0.7
~
. ----4 I
__ J ~
0.5
go
0.4
Q:
03 0.2
line).
01
12
16
20
24
16
20
24
Weeks
1.0
C
0.9 08 07 ~
~ 0.6
e
005
;:
8.
£.
0.4 j----
0.3 0.2 0.1
12
Weeks
leagues (14)found 2 of 19 patients (11 070) to have vocal cord granulomas 16wk after extubation. Burns and coworkers (15) detected no laryngeal granulomas in 117 patients by 12wk after extubation. Lindholm (2) carefully studied 69 adults by laryngoscopy every 15days following extubation. Of these patients, 43 healed the larynx by primary reepithelialization within 4 wk. The other 26 patients developed initial granulomas at the site of laryngeal ulcers, but in 23 patients these granulomas spontaneously resolved by a median time of 9 wk. Two patients re-
quired surgical removal of granulomas, and one patient had surgical correction of a posterior commissure fibrous band. These studies agreed that laryngeal healing is usually completed by 9 to 16 wk after extubation with a low rate of scarring or persistent granulomas. A limiting factor in all these studies, however, was failure to adequately define the influence of tracheostomy on resolution of laryngeal injury. Both Burns and coworkers (15) and Lindholm (2) studied patients who underwent tracheostomy following TLI, but did not separately report the rate
of abnormal laryngeal healing in these patients. Kastonas and colleagues(14) did not include patients receiving tracheostomy in their study. Whited studied laryngeal healing in 200 patients who had TLI for 2 to 24 days (3,4). The majority of these patients had a normal larynx by 3 wk after extubation, with healing seemingly faster with shorter duration TLI. Twelve (6%) of these patients developed laryngeal scarring by 16 wk following extubation. In 11 patients scarring was observed in the interarytenoid plane and inferiorly along the internal face of the cricoid. The other patient developed fibrous bands between the vocal cords. Although all 12 of these patients had undergone tracheostomy in addition to TLI, Whited believed that tracheostomy was not an important variable promoting abnormal laryngeal healing. Instead, he speculated that the severity of laryngeal injury at extubation was the critical factor influencing the development of laryngeal scarring. This study has avoided the limitations of previous studies. Patients experienced prolonged TLI before extubation or tracheostomy, reflecting general practice in present-day intensive care units. Followup was complete for all patients and direct visual examination of the larynx was performed in a prospective, systematic manner. This allows the true rate of laryngeal healing to be expressed. Patients undergoing tracheostomy were included to allow evaluation of this procedure's effect on resolution of laryngeal injury. Unlike Whited, we found no cases of laryngeal scarring and a small rate (3 of 51, or 6%) of persistent laryngeal granulomas. Healing was usually complete by 8 to 12wk. These results are similar to those described by others (2, 14, 15). These observations have important clinical implications. Because severity of laryngeal injury at extubation did not influence either the rate or manner of
364
laryngeal healing, laryngoscopy findings at extubation should not be used to determine whether tracheostomy is needed to minimize the development of future laryngeal pathology. Laryngoscopy should be considered, however, in patients who have hoarseness persisting beyond the median time to resolution of laryngeal pathology, which was 4 wk in this study. Surgically correctable problems, such as laryngeal granulomas, may be found in these cases. The finding that duration of TLI did not influence laryngeal healing, in conjunction with our previous observation that duration of TLI does not appear to affect severity of laryngeal injury at extubation, disputes the commonly held viewpoint that secondary tracheostomy is advisable in patients requiring prolonged TLI for mechanical ventilation to protect the larynx, Unfortunately, this study does not provide insights into what variables promote abnormal laryngeal healing. Besides the factors detailed in this study, no other clinically obvious common theme, such as aspiration, poor nutrition, or type of underlying disease, was found in the patients who developed laryngeal granulomas. Presumably, individual variability in wound healing, a poorly defined but
GENE L. COLICE
well-recognized concept, determined whether laryngeal injury healed by primary reepithelialization. However, this study reassures the clinician that abnormal laryngeal healing, by either scarring or granuloma formation, occurs much less often and with less morbidity than recently reported by others (3, 4). In conclusion, laryngeal injury at extubation is common, but healing by primary reepithelialization usually occurred within 8 wk. Few patients developed laryngeal granulomas, and no cases of laryngeal scarring wereobserved. Tracheostomy, TLI lasting more than 10 days, and severelaryngeal injury at extubation did not influence laryngeal healing. Acknowledgment The author appreciates the advice on statistical evaluation of the data provided by Therese Stukel, Ph.D. and Bradley Dain, M.S., Biostatistical Shared Service; Darmouth Medical School.
References 1. Colice GL, Stukel TA, Dain B. Laryngeal complications of prolonged intubation. Chest 1989; 96:877-84.
2. Lindholm CEo Prolonged endotracheal intubation. Acta Anaesthesiol Scand 1969; 33:1-131. 3. Whited RE. Posterior commissurestenosispost
long-term intubation. Laryngoscope 1983; 93: 1314-8.
4. Whited RE. A prospective study of laryngotracheal sequelae in longer-term intubation. Laryngoscope 1984; 94:367-77. 5. Kalbfleisch ID, Prentice RL. The statistical analysisof failuretime data. NewYork:John Wiley and Sons, 1980. 6. Cox DR. Regression models and life tables. 1 R Stat SOC IB] 1972; 34:187-220. 7. Kambic V, Radsel Z. Intubation lesion of the larynx. Br 1 Anaesth 1978; 50:587-90. 8. Arner 0, Diamant M. Respiratory-tract lesions following intratracheal anaesthesia. Acta Chir Scand 1951; 101:75-84. 9. Aass AS. Complications to tracheostomy and long-term intubation: a follow-up study. Acta Anaesthesiol Scand 1975; 19:127-33. 10. Stauffer Jl., Olson DE, Petty TL. Complications and consequences of endotracheal intubation and tracheostomy. Am 1 Med 1981; 70:65-76. 11. Tonkin IP, Harrison GA. The effect on the larynx of prolonged endotracheal intubation. Med 1 Aust 1%6; 2:581-7. 12. Dixon TC, Sando MIW, Bolton 1M, Gilligan lE. A report of 342 casesof prolonged endotracheal intubation. Med 1 Aust 1968; 2:529-33. 13. Dunham CM, LaMonica C. Prolonged intubation in the trauma patient. 1 Trauma 1984; 24:120-4. 14. Kastanos N, Miro RE, PerezAM, etal. Laryn-
gotracheal injury due to endotracheal intubation. Crit Care Med 1983; 11:362-7. 15. Burns HP, Dayal VS, Scott A, et al. Laryngotracheal trauma: observations on its pathologenesis and its prevention following prolonged orotracheal intubation in the adult. Laryngoscope 1979; 89:1316-24.
