The Role of Inhalation Injury in
Burn Trauma
A Canadian Experience
EDWARD E. TREDGET, M.Sc., M.D.,* HEATHER A. SHANKOWSKY, R.N.,* TERRY V. TAERUM, PH.D.,t GERALD L. MOYSA, M.D.,* and JOHN D. M. ALTON, M.D.*
From 1977 to 1987, 1705 thermally injured patients were admitted to the Firefighters' Burn Center at the University of Alberta Hospitals. Thirteen hundred forty-four were male (78.8%) and 361 were female (21.2%), with a mean total burn surface area (TBSA) of 15.1 (SEM ± 0.4%) and a range of 1% to 99% TBSA. Sixteen hundred thirty-five patients survived to be discharged from hospital, with an overall survival rate of 95.9%. One hundred twenty-four burn patients (7.3%) suffered concomitant inhalation injury diagnosed by bronchoscopy. Patients with inhalation injury suffered from larger TBSA (39.7% ± 2.8% versus 12.2% ± 0.3%; p < 0.01) than those without inhalation injury. Inhalation injury increased the number of deaths from burn injury (34.7% versus 1.7%; p < 0.01) independent of age and TBSA. Inhalation injury was associated with a threefold prolongation of hospital stay (23.7 ± 0.7 versus 74.4 ± 6.2 days; p < 0.01) and was independent of age and TBSA. Multifactorial probit analysis was performed for both inhalation- and noninhalationinjured burned patients to allow TBSA and age adjusted rates of mortality for the burn population presented. The maximum detrimental effects of inhalation injury in burn patient outcome occurred when it coexisted with moderate (15% to 29% TBSA) to large (30% to 69% TBSA) thermal injuries. These data demonstrate that inhalation injury is an important comorbid factor in burn injury that increases the number of deaths substantially. Most importantly such injuries also independently prolong the duration of hospitalization in a highly unpredictable fashion as compared to patients with cutaneous burns only. As such our data illustrate the extreme importance of inhalation injury as a comorbid factor following thermal injury and reveal the present limitations for accurate quantification of the magnitude of respiratory tract injury accompanying thermal trauma.
D
~ URING THE LAST decade, improvements in resuscitation of bum-injured and other trauma patients has led to increased incidence of sur-
Supported by the Alberta Heritage Foundation for Medical Research and by the University of Alberta Hospitals Firefighters' Burn Treatment Fund. Address reprint requests to Edward E. Tredget, M.D., Director, Firefighters' Burn Treatment Unit, 2D3.82, Walter McKenzie Centre, University of Alberta, Edmonton, Alberta, Canada T6G 2B7. Accepted for publication December 27, 1989.
From the Firefighters' Burn Treatment Unit,* Division of Plastic Surgery, Department of Surgery, and University Computing Systems,t University of Alberta, Edmonton, Alberta, Canada
vival from bum shock after major thermal injury; however sepsis and pulmonary failure remain important and principal causes ofmortality and morbidity following thermal
injury.' Thermally damaged skin as well as inhaled products of combustion are known to activate neutrophils, in part through the alternate pathway ofthe complement system, subsequently damaging pulmonary tissues through release of proteases and free oxygen radicals. 1-3 Similarly stimulation of pulmonary macrophages following respiratory tract injury produces prostanoids and cytokines that further contribute to the pathophysiology of inhalation injury and the morbidity and mortality of the host by mechanisms incompletely understood.4 Although this increase in burn mortality rate has been quantitated up to 60% beyond that expected in patients with cutaneous burns only,5 the impact of inhalation injury on morbidity, length of hospitalization, and cost of care has not been thoroughly addressed. The following study therefore was undertaken to investigate further the role of inhalation injury on both mortality and morbidity following thermal injury and to establish the overall duration of hospitalization and rate of recovery following thermal injury in a burn center. functioning within the Canadian health care system. Materials and Methods The records of 1713 patients were reviewed during the 11-year period from January 1, 1977 to December 31, 1987. Epidemiologic, demographic, and therapeutic patient data were recorded in a computerized burn registry.
720
ROLE OF INHALATION INJURY IN BURN TRAUMA
Vol. 212 * No. 6
This registry was on a personal computer that was linked, via network, to an Amdahl 5870 mainframe computer (Sunnyvale, CA) operating under the Michigan Terminal System (MTS) (University of Michigan, Ann Arbor, MI). Of the 1713 records reviewed, eight patients with extreme injuries who received compassionate care only were excluded. Patients cared for were referred to the burn center from the central and northern regions of the province of Alberta and the North West Territories of Canada and were resuscitated uniformly using the Parkland formnula.6 Burn wounds were treated with topical application of silver sulfadiazine every 12 hours until excision and grafting of deep second- and third-degree wounds were commenced, usually within the first postburn week. Inhalation injury was suspected in patients with a history of impaired level of consciousness, injury occurring in closed space or structural fires, and in those patients sustaining facial burns. The diagnosis of inhalation injury was confirmed by the presence of mucosal erythema, edema, or ulceration and submucosal hemorrhage found on bronchoscopy normally performed within the first 24 hours after burn.7 The charts of all patients requiring ven-
721
Distribution of Burn Size 1000
800
600 a.)
a)
0)
LL 400
200
Distribution of Age 0 oi
30r
W0
,,0
,oY
"'% \cbo
FIG. 2. Distribution of burn sizes (TBSA %). Overall TBSA for 1705 patients was 15.1% ± 0.4%.
25F
tilatory support within the first 7 days after injury also were reviewed for the possibility of inhalation injury.8 Although arterial blood gases and carboxyhemoglobin were measured and chest x-rays performed, they were unreliable for the establishment of the presence of inhalation injury. Pulmonary function studies and '33Xenon lung scans were used only in a small number of patients and were not regularly performed for practical reasons. Statistical analysis was performed using the MANOVA, CROSSTABS, UANOVA, and PROBIT procedures in SPSS-X® (SPSS Statistical Program, Chicago, IL) and the logistical regression procedure in BMDP® (BMDP Statistical Software, Inc., Los Angeles, CA). The relative effects of burn size, age, and inhalation injury on death were analyzed with multifactorial probit analysis.9"'0
2010)
15a)
10
,q ,0
"0 (%)¼bl@@4 OoTB TBSA (%)
~
5h
0
,v
,vAge
(years)
Age (years) FIG. 1. Distribution of ages for 1705 patients included in the study. Mean age, 26.2 ± 0.5 years.
Results Of the 1705 patients in the study, 1344 were male (78.8%) and 361 were female (21.2%), of which 1635 patients survived, giving an overall survival rate of 95.9%. The frequency distribution of age, as shown in Figure 1,
TREDGET
722
AND OTHERS
60
TABLE 1. Comparison of Descriptive Statistics from Four Burn Centers
Statistic
UAH
MGH"
Brooke'
Salt Lake City12
Number Mean age (years) Mean TBSA (%) Mean LOS (days) % Inhalation Mortality rate (%)
1705 26.2 15.1 26.2 7.3 4.1
619 39.0 17.0 23.0 N/A 7.0
1058 31.0 23.7 N/A 24.6 22.6
1458 24.4 19.0 14.7 7.0 8.0
demonstrates that the largest group of patients within our population was between the age of 20 and 29 years (30%), 86% of whom were male. Of the 1705 patients, 36% were in the pediatric or adolescent age range of 0 to 19 years, 5 1% adults were in the working years of life and only 7% of the patients were 60 years or older. The mean age of the population studied is 26.2 ± 0.5 years (range, 0 to 90 years). Most burn injuries in our study occurred at home (48%), with 34% occurring at work and 19% at other locations apart from the home or work place. Sixty per cent of these injuries were thermal burns, 19% were due to scalds, 7% to electrical injuries, and 14% were due to other injuries (including tar, grease, steam, friction, and other chemicals). The mean size of burn injury was 15.1% ± 0.4% (range, 1% to 100%), with 26.2% ofthe injuries in the range from Mortality-
A n% 'IV
Ann. Surg. December 1990
Mortality
50
40
=
30
0
10
0
TBSA (%) FIG. 4. Mortality in burn patients versus TBSA. The overall mortality rate was 4.1%.
20% to 100% of the total body surface and only 3.2% of the injuries were massive (70% or more) (Fig. 2). The mean age of the patient population and burn size does not differ significantly in each of the years of the study. The overall mortality rate in this study is 4.1%, which 30 compares favorably with other large studies (Table i ).5,8"11,12 Figures 3 and 4 illustrate that the number of burn deaths increases exponentially with increasing age of the patient and size of the burn injury. One hundred twenty-four individuals of the 1705 patients studied suffered concomitant inhalation injury 20 based on the criteria established earlier.8 The mean age of the patients who suffered both inhalation injury and thermal burns was 30.1 ± 1.4 years, somewhat higher than for the patients with only cutaneous burns (25.2 ± 0.5 years; p > 0.05, not significant [NS]) (Fig. 5). However, 10 for each grouping of TBSA, that is small (1% to 14%), moderate (15% to 29%), large (30% to 69%), and massive (70%), no statistically significant difference in age of the patients existed between those patients with inhalation and burns and those with burns only, thus allowing further 0 _ . X comparison of the mortality of bums with and without 0'9 0.\9 11"~'15 '~10 o,fb9 0,W9 tpr%,g9 o,9 0 bZ' _inhalation injury, without the confounding effects of age. Age (years) In each age grouping of burned patients, the size of burn was substantially larger for patients with inhalational Fw G 3. Mortality in burn patients versus age. The overall mortity rate injury. The overall burn size was 39.7% ± 2.8% for patients waLS 4.1%. -
-
-
.*
40 r
723
ROLE OF INHALATION INJURY IN BURN TRAUMA
Vol. 212 . No. 6
Mean Age (With/Without Inhalation Injury)
with patients with no respiratory injury (Fig. 10). From the step-wise logistic regression, these data can be represented by the predictive equation:
as contrasted
1
35-
where: Inhalation Injury
au
6.00 - 6.90 X 10-2 (TBSA) - 8.08 X 10-8 (A)3, for the whole population and u = 4.99 - 4.26 X 10-2 (TBSA) - 1.30 (I) - 8.65 X 10-6 (A)3 - 2.12 X l0-4 (TBSA) (A), for those patients with and without inhalation injury, where p is the probability of survival, I = 1 with inhalation injury and 1 if no inhalation injury is present, TBSA is total body surface area burned, and A is the age of the patient. Finally morbidity of burn injury, in part depicted by length of stay (LOS) in hospital, is shown for all patients in Figure 11 and illustrates a gradual decline in the last 7 years of the study period. However, when one examines u =
30
a 0)
c
20
15
eu + eu
1-14
15-29
30-69
70+
TBSA (%) Mean TBSA
FIG. 5. Mean age versus %TBSA for patients with and without inhalation injury. The oveall age of the patients without an associated inhalation injury was 25.2 ± 0.5 years compared to 30.1 ± 1.4 years for those patients with an inhalation injury.
(With/Without Inhalation Injury) 80
with inhalation injury compared to 12.2% ± 0.3% for those who had burn injury only (p < 0.01) (Fig. 6). By analysis of variance, a substantial and significant difference in mortality rate was evident for patients in various age groupings (Fig. 7) and for each size of burn (Fig. 8) (1.7% mortality rate with no inhalation injury compared to 34.7% mortality rate with inhalation injury; p 70%) in burn patients with and without inhalation injury.
N\
80
0
20
40 60 Age (years)
80
100
30-89% TBSA (With/Without Inhalation Inhuy)
No
Inhalation Injury
N\ N\ Inhlation
Injury\
Ns 0
20
40 60 Age (years)
80
100
;70 % TBSA (With/Without Inhalation Injury)
100 80 -a 60 60
40 inhdaton k*8
20 0
20
;
40 60 Age (years)
80
reverse the otherwise favorable survival rates (Fig. 7) and prolongs the length of stay in hospital (Fig. 13). Both these factors are probably due to the inherent difficulties in young children of maintaining good tracheobronchial toilet to ensure adequate ventilation through smaller airways and narrower endotracheal tubes. We have demonstrated, as have others,5,8,10,'2,'8 that inhalation injury contributes to mortality in burns, but its influence is predominantly on patients with moderate and large burns who otherwise have favorable prognosis
100
40 60 Age (years)
without inhalation injury (Fig. 9). This is best demonstrated by shift in the slope of declining survival with inhalation injury in this range of burn size independent of age, as demonstrated in .the three-dimensional probit analyses as compared to those patients with burn injury only. Furthermore, although most attention has been directed to mortality caused by inhalation injury in burn patients, the morbidity of inhalation injury and its effect on length of stay in hospital has not been realized totally. Through analysis of covariance, our data suggests that
Survival Without Inhalation Injury
£I 6.1
FIGS. 10 A and B. Three dimensional plot of probability of survival with inhalation injury (B).
following burn injury in patients without associated inhalation injury (A) and
726
TREDGET AND OTHERS
Length of Stay
50
401 (0
, 30
I~~~~~~~~~ "III--I I'-----I---"'I
co
cn
0
-j20 c
0
10
100
80 0
cU V
co 0
60
-J
X 40
Ann. Surg. December 1990
severe form of injury than detectable by bronchoscopy. In the presence of bronchoscopic pathology, this investigation added little to the diagnosis and, in their assessment, did not warrant the practical difficulties encountered in obtaining the results.5 In laboratory models of inhalation injury, Zawacki et al.2' and others22'23 found that severity of inhalation injury was a function of both the size of burn and the amount of smoke inhaled, and within limits carboxyhemoglobin levels were proportional to severity of injury. However the usefulness of carboxyhemoglobin levels as a predictor of the severity of inhalation injury has not been substantiated in burn patients,'8 in keeping with our own experience. Thus few clinical indices exist as yet that allow quantitative analysis of the magnitude of pulmonary injury after smoke inhalation. Early identification and quantification of the magnitude of injury in our inhalation-injured patients would allow identification of those patients who develop increased deadspace ventilation, prolonged requirements for mechanical ventilation, and are at risk of developing chronic pulmonary insufficiency, physiologic abnormalities that were documented in a subsegment of our inhalation-injured burn patients.24 As
20 77
78
79
80
81
82
83
84
85
86
Mean Length of Stay (With/Without Inhalation Injury)
87
Admission Year
FIG. 11. Length of stay in days versus year of admission. Top graph includes all admissions and the bottom graph shows the same population divided into those burn patients with and without inhalation injury. Data shown are the mean and the SEM for each year. The mean length of stay for the total number of patients was 26.2 ± 0.8 days. The mean length of stay for those patients without an associated inhalation injury was 23.7 ± 0.7 days compared to 74.4 ± 6.2 days for those patients with an inhalation injury (p < 0.01).
duration of stay in hospital is increased threefold on average for those surviving patients with inhalation injury and this prolongation of the morbid period is independent of the age (Fig. 13) and the size of the burn injury (Fig. 12). More importantly the duration of hospitalization in our center can be predicted within a very narrow range if patients have no respiratory tract involvement with their cutaneous burns, but when inhalation injury coexists, the mean length of stay is substantially longer and highly variable, as expressed by the wide standard error of the mean for each measurement (Figure 11). These factors become important in overall assessment of cost of care and make resource and manpower planning difficult.'9 They also illustrate the difficulty in assessing the severity ofthe respiratory tract 'burn' and the limitations of bronchoscopic assessment only in inhalation injury.20 Although Shirani et al.5 used 133Xenon lung scanning to detect respiratory tract damage in the absence of abnormalities at bronchoscopy, such cases represented a less
140
1201100
V)
80
0 c
co,
a)
60 |Inhalation Injury
40 20
0
1-14
30-69 TBSA (%)
15-29
70+
FIG. 12. Mean length of stay (LOS) versus %TBSA for small, moderate, large, and massive burns. Data shown are the mean for each group and the SEM. The mean LOS for those patients without an associated inhalation injury was 23.7 ± 0.7 days compared to 74.4 ± 6.2 days for those patients with an inhalation injury (p < 0.01).
Vol. 212 * No.6
ROLE OF INHALATION INJURY IN BURN TRAUMA
Mean Length of Stay (With/Without Inhalation Injury)
240 220
References
* -
200 180 .
160 .
3C10,
140
-
T
,
3J12080
Ko a -
Inhalation Injury ,
j
727
:
I
60
Age (years) FIG. 13. Mean length of stay (LOS) versus age for burn patients with and without inhalation injury. Data shown are the mean for each age grouping and the SEM. The mean LOS for those patients without an associated inhalation injury was 23.7 ± 0.7 days compared to 74.4 ± 6.2 days for those patients without an inhalation injury (p < 0.01).
the understanding of the pathophysiology of this injury increases, analysis of free radicals, prostanoids, or cytotoxic lymphokines in the pulmonary tissues'13'25 and the interaction of the burn wound with pulmonary function26 will allow increased understanding of both the severity of inhalation injury in burn patients and its treatment. Our data further documents that inhalation injury increases burn mortality independent of patient age and burn size. In surviving thermally injured adults and children, hospitalization is prolonged substantially by inhalation injury, again independent of the age and TBSA in a highly variable and as yet unpredictable fashion.
Acknowledgments The authors thank the residents, nurses, physiotherapists, occupational therapists, dieticians, staff physicians, and all others on the burn team for the excellent care they gave to the patients involved in this study.
1. Demling RH. Bums. N Engl J Med 1985; 313:1389-1393. 2. Traber EL, Linares HA, Hemdon DN. The pathophysiology of inhalation injury-a review. Bums 1988; 14:357-364. 3. Kinsella J. Smoke inhalation. Bums 1988; 14:269-279. 4. Gemmell CG, Pollok AJ, McMillan F, et al. Structural and functional changes in alveolar macrophages following the exposure of fire victims to smoke. Eur J Clin Invest 1987; 17:321-325. 5. Shirani KZ, Pruitt BA, Mason AD. The influence of inhalation injury and pneumonia on bum mortality. Ann Surg 1987; 20:82-87. 6. Baxter CR, Shires T. Physiologic response to crystalloid resuscitation of severe bums. Ann NY Acad Sci 1968; 150:874-880. 7. Hunt JL, Agee RN, Pruitt BA. Fiberoptic bronchoscopy in acute inhalation injury. J Trauma 1975; 15:641-649. 8. Thompson PB, Hemdon DN, Traber DL, et al. Effect on mortality of inhalation injury. J Trauma 1986; 26:163-165. 9. Margosches EH, Roi LD, Flora JD. The statistical analysis of bum patient data: a historical review. Bums 1978; 5:43-53. 10. Zawacki BE, Azen SP, Imbus SH, Chang YC. Multifactorial probit analysis of mortality in bumed patients. Ann Surg 1979; 189: 1 5. 11. Tompkins RG, Burke JF, Schoenfeld DA, et al. Prompt eschar excision: a treatment system contributing to reduced bum mortality. Ann Surg 1986; 194:272-281. 12. Merrell SW, Saffle JR, Sullivan JJ, et al. Increased survival after major thermal injury. Am J Surg 1987; 154:623-627. 13. Tompkins RG, Remensnyder JP, Burke JF, et al. Significant reductions in mortality for children with bum injuries through the use of prompt eschar excision. Ann Surg 1988; 208:33-41. 14. Stem M, Waisbren BA. A method by which bum units may compare their results with a base line curve. Surg Gynecol Obstet 1976; 142:230-234. 15. Flora JD. A method for comparing survival of bum patients to a standard survival curve. J Trauma 1976; 18:701-705. 16. Iglehart JK. Canada's health care system. N Engl J Med 1986; 315: 202-208. 17. Cox DR. Analysis of Binary Data. London: Chapman and Hall, 1970, pp 14-29. 18. Clark WR, Bonaventura M, Myers W. Smoke inhalation and airway management at a regional bum unit: 1974-1983. J Bum Care Rehab 1989; 10:52-62. 19. Curreri PW, Luterman A, Braun D, Shires GT. Bum injury: analysis of survival and hospitalization time for 937 patients. Ann Surg 1980; 192:472-478. 20. Bingham HG, Gallagher TJ, Powell MD. Early bronchoscopy as a predictor of ventilatory support for bumed patients. J Trauma 1987; 27:1286-1289. 21. Zawacki BE, Jung RC, Joyce J, Rincon E. Smoke, bums and the natural history of inhalation injury in fire victims: a correlation ofexperimental and clinical data. Ann Surg 1977; 185:100-110. 22. Hemdon DN, Traber DL, Niehaus GD, et al. The pathophysiology of smoke inhalation injury in a sheep model. J Trauma 1984; 24:1044-1051. 23. Shimazu T, Yukioka T, Hubbard GB, et al. A dose-responsive model of smoke inhalation injury. Ann Surg 1987; 206:89-97. 24. Stollery DE, Jones RL, King EG. Deadspace ventilation: a significant factor in respiratory failure after thermal inhalation. Crit Care Med 1987; 15:260-261. 25. Clark CJ, Pollock AJ, Reid WH, et al. Role of pulmonary alveolar macrophage activation in acute lung injury after bums and smoke inhalation. Lancet 1988; ii:872-877. 26. Demling RH, Will JA, Belzer FO. Effect of major thermal injury on the pulmonary microcirculation. Surgery 1978; 83:746-751.
Burn Trauma
A Canadian Experience
EDWARD E. TREDGET, M.Sc., M.D.,* HEATHER A. SHANKOWSKY, R.N.,* TERRY V. TAERUM, PH.D.,t GERALD L. MOYSA, M.D.,* and JOHN D. M. ALTON, M.D.*
From 1977 to 1987, 1705 thermally injured patients were admitted to the Firefighters' Burn Center at the University of Alberta Hospitals. Thirteen hundred forty-four were male (78.8%) and 361 were female (21.2%), with a mean total burn surface area (TBSA) of 15.1 (SEM ± 0.4%) and a range of 1% to 99% TBSA. Sixteen hundred thirty-five patients survived to be discharged from hospital, with an overall survival rate of 95.9%. One hundred twenty-four burn patients (7.3%) suffered concomitant inhalation injury diagnosed by bronchoscopy. Patients with inhalation injury suffered from larger TBSA (39.7% ± 2.8% versus 12.2% ± 0.3%; p < 0.01) than those without inhalation injury. Inhalation injury increased the number of deaths from burn injury (34.7% versus 1.7%; p < 0.01) independent of age and TBSA. Inhalation injury was associated with a threefold prolongation of hospital stay (23.7 ± 0.7 versus 74.4 ± 6.2 days; p < 0.01) and was independent of age and TBSA. Multifactorial probit analysis was performed for both inhalation- and noninhalationinjured burned patients to allow TBSA and age adjusted rates of mortality for the burn population presented. The maximum detrimental effects of inhalation injury in burn patient outcome occurred when it coexisted with moderate (15% to 29% TBSA) to large (30% to 69% TBSA) thermal injuries. These data demonstrate that inhalation injury is an important comorbid factor in burn injury that increases the number of deaths substantially. Most importantly such injuries also independently prolong the duration of hospitalization in a highly unpredictable fashion as compared to patients with cutaneous burns only. As such our data illustrate the extreme importance of inhalation injury as a comorbid factor following thermal injury and reveal the present limitations for accurate quantification of the magnitude of respiratory tract injury accompanying thermal trauma.
D
~ URING THE LAST decade, improvements in resuscitation of bum-injured and other trauma patients has led to increased incidence of sur-
Supported by the Alberta Heritage Foundation for Medical Research and by the University of Alberta Hospitals Firefighters' Burn Treatment Fund. Address reprint requests to Edward E. Tredget, M.D., Director, Firefighters' Burn Treatment Unit, 2D3.82, Walter McKenzie Centre, University of Alberta, Edmonton, Alberta, Canada T6G 2B7. Accepted for publication December 27, 1989.
From the Firefighters' Burn Treatment Unit,* Division of Plastic Surgery, Department of Surgery, and University Computing Systems,t University of Alberta, Edmonton, Alberta, Canada
vival from bum shock after major thermal injury; however sepsis and pulmonary failure remain important and principal causes ofmortality and morbidity following thermal
injury.' Thermally damaged skin as well as inhaled products of combustion are known to activate neutrophils, in part through the alternate pathway ofthe complement system, subsequently damaging pulmonary tissues through release of proteases and free oxygen radicals. 1-3 Similarly stimulation of pulmonary macrophages following respiratory tract injury produces prostanoids and cytokines that further contribute to the pathophysiology of inhalation injury and the morbidity and mortality of the host by mechanisms incompletely understood.4 Although this increase in burn mortality rate has been quantitated up to 60% beyond that expected in patients with cutaneous burns only,5 the impact of inhalation injury on morbidity, length of hospitalization, and cost of care has not been thoroughly addressed. The following study therefore was undertaken to investigate further the role of inhalation injury on both mortality and morbidity following thermal injury and to establish the overall duration of hospitalization and rate of recovery following thermal injury in a burn center. functioning within the Canadian health care system. Materials and Methods The records of 1713 patients were reviewed during the 11-year period from January 1, 1977 to December 31, 1987. Epidemiologic, demographic, and therapeutic patient data were recorded in a computerized burn registry.
720
ROLE OF INHALATION INJURY IN BURN TRAUMA
Vol. 212 * No. 6
This registry was on a personal computer that was linked, via network, to an Amdahl 5870 mainframe computer (Sunnyvale, CA) operating under the Michigan Terminal System (MTS) (University of Michigan, Ann Arbor, MI). Of the 1713 records reviewed, eight patients with extreme injuries who received compassionate care only were excluded. Patients cared for were referred to the burn center from the central and northern regions of the province of Alberta and the North West Territories of Canada and were resuscitated uniformly using the Parkland formnula.6 Burn wounds were treated with topical application of silver sulfadiazine every 12 hours until excision and grafting of deep second- and third-degree wounds were commenced, usually within the first postburn week. Inhalation injury was suspected in patients with a history of impaired level of consciousness, injury occurring in closed space or structural fires, and in those patients sustaining facial burns. The diagnosis of inhalation injury was confirmed by the presence of mucosal erythema, edema, or ulceration and submucosal hemorrhage found on bronchoscopy normally performed within the first 24 hours after burn.7 The charts of all patients requiring ven-
721
Distribution of Burn Size 1000
800
600 a.)
a)
0)
LL 400
200
Distribution of Age 0 oi
30r
W0
,,0
,oY
"'% \cbo
FIG. 2. Distribution of burn sizes (TBSA %). Overall TBSA for 1705 patients was 15.1% ± 0.4%.
25F
tilatory support within the first 7 days after injury also were reviewed for the possibility of inhalation injury.8 Although arterial blood gases and carboxyhemoglobin were measured and chest x-rays performed, they were unreliable for the establishment of the presence of inhalation injury. Pulmonary function studies and '33Xenon lung scans were used only in a small number of patients and were not regularly performed for practical reasons. Statistical analysis was performed using the MANOVA, CROSSTABS, UANOVA, and PROBIT procedures in SPSS-X® (SPSS Statistical Program, Chicago, IL) and the logistical regression procedure in BMDP® (BMDP Statistical Software, Inc., Los Angeles, CA). The relative effects of burn size, age, and inhalation injury on death were analyzed with multifactorial probit analysis.9"'0
2010)
15a)
10
,q ,0
"0 (%)¼bl@@4 OoTB TBSA (%)
~
5h
0
,v
,vAge
(years)
Age (years) FIG. 1. Distribution of ages for 1705 patients included in the study. Mean age, 26.2 ± 0.5 years.
Results Of the 1705 patients in the study, 1344 were male (78.8%) and 361 were female (21.2%), of which 1635 patients survived, giving an overall survival rate of 95.9%. The frequency distribution of age, as shown in Figure 1,
TREDGET
722
AND OTHERS
60
TABLE 1. Comparison of Descriptive Statistics from Four Burn Centers
Statistic
UAH
MGH"
Brooke'
Salt Lake City12
Number Mean age (years) Mean TBSA (%) Mean LOS (days) % Inhalation Mortality rate (%)
1705 26.2 15.1 26.2 7.3 4.1
619 39.0 17.0 23.0 N/A 7.0
1058 31.0 23.7 N/A 24.6 22.6
1458 24.4 19.0 14.7 7.0 8.0
demonstrates that the largest group of patients within our population was between the age of 20 and 29 years (30%), 86% of whom were male. Of the 1705 patients, 36% were in the pediatric or adolescent age range of 0 to 19 years, 5 1% adults were in the working years of life and only 7% of the patients were 60 years or older. The mean age of the population studied is 26.2 ± 0.5 years (range, 0 to 90 years). Most burn injuries in our study occurred at home (48%), with 34% occurring at work and 19% at other locations apart from the home or work place. Sixty per cent of these injuries were thermal burns, 19% were due to scalds, 7% to electrical injuries, and 14% were due to other injuries (including tar, grease, steam, friction, and other chemicals). The mean size of burn injury was 15.1% ± 0.4% (range, 1% to 100%), with 26.2% ofthe injuries in the range from Mortality-
A n% 'IV
Ann. Surg. December 1990
Mortality
50
40
=
30
0
10
0
TBSA (%) FIG. 4. Mortality in burn patients versus TBSA. The overall mortality rate was 4.1%.
20% to 100% of the total body surface and only 3.2% of the injuries were massive (70% or more) (Fig. 2). The mean age of the patient population and burn size does not differ significantly in each of the years of the study. The overall mortality rate in this study is 4.1%, which 30 compares favorably with other large studies (Table i ).5,8"11,12 Figures 3 and 4 illustrate that the number of burn deaths increases exponentially with increasing age of the patient and size of the burn injury. One hundred twenty-four individuals of the 1705 patients studied suffered concomitant inhalation injury 20 based on the criteria established earlier.8 The mean age of the patients who suffered both inhalation injury and thermal burns was 30.1 ± 1.4 years, somewhat higher than for the patients with only cutaneous burns (25.2 ± 0.5 years; p > 0.05, not significant [NS]) (Fig. 5). However, 10 for each grouping of TBSA, that is small (1% to 14%), moderate (15% to 29%), large (30% to 69%), and massive (70%), no statistically significant difference in age of the patients existed between those patients with inhalation and burns and those with burns only, thus allowing further 0 _ . X comparison of the mortality of bums with and without 0'9 0.\9 11"~'15 '~10 o,fb9 0,W9 tpr%,g9 o,9 0 bZ' _inhalation injury, without the confounding effects of age. Age (years) In each age grouping of burned patients, the size of burn was substantially larger for patients with inhalational Fw G 3. Mortality in burn patients versus age. The overall mortity rate injury. The overall burn size was 39.7% ± 2.8% for patients waLS 4.1%. -
-
-
.*
40 r
723
ROLE OF INHALATION INJURY IN BURN TRAUMA
Vol. 212 . No. 6
Mean Age (With/Without Inhalation Injury)
with patients with no respiratory injury (Fig. 10). From the step-wise logistic regression, these data can be represented by the predictive equation:
as contrasted
1
35-
where: Inhalation Injury
au
6.00 - 6.90 X 10-2 (TBSA) - 8.08 X 10-8 (A)3, for the whole population and u = 4.99 - 4.26 X 10-2 (TBSA) - 1.30 (I) - 8.65 X 10-6 (A)3 - 2.12 X l0-4 (TBSA) (A), for those patients with and without inhalation injury, where p is the probability of survival, I = 1 with inhalation injury and 1 if no inhalation injury is present, TBSA is total body surface area burned, and A is the age of the patient. Finally morbidity of burn injury, in part depicted by length of stay (LOS) in hospital, is shown for all patients in Figure 11 and illustrates a gradual decline in the last 7 years of the study period. However, when one examines u =
30
a 0)
c
20
15
eu + eu
1-14
15-29
30-69
70+
TBSA (%) Mean TBSA
FIG. 5. Mean age versus %TBSA for patients with and without inhalation injury. The oveall age of the patients without an associated inhalation injury was 25.2 ± 0.5 years compared to 30.1 ± 1.4 years for those patients with an inhalation injury.
(With/Without Inhalation Injury) 80
with inhalation injury compared to 12.2% ± 0.3% for those who had burn injury only (p < 0.01) (Fig. 6). By analysis of variance, a substantial and significant difference in mortality rate was evident for patients in various age groupings (Fig. 7) and for each size of burn (Fig. 8) (1.7% mortality rate with no inhalation injury compared to 34.7% mortality rate with inhalation injury; p 70%) in burn patients with and without inhalation injury.
N\
80
0
20
40 60 Age (years)
80
100
30-89% TBSA (With/Without Inhalation Inhuy)
No
Inhalation Injury
N\ N\ Inhlation
Injury\
Ns 0
20
40 60 Age (years)
80
100
;70 % TBSA (With/Without Inhalation Injury)
100 80 -a 60 60
40 inhdaton k*8
20 0
20
;
40 60 Age (years)
80
reverse the otherwise favorable survival rates (Fig. 7) and prolongs the length of stay in hospital (Fig. 13). Both these factors are probably due to the inherent difficulties in young children of maintaining good tracheobronchial toilet to ensure adequate ventilation through smaller airways and narrower endotracheal tubes. We have demonstrated, as have others,5,8,10,'2,'8 that inhalation injury contributes to mortality in burns, but its influence is predominantly on patients with moderate and large burns who otherwise have favorable prognosis
100
40 60 Age (years)
without inhalation injury (Fig. 9). This is best demonstrated by shift in the slope of declining survival with inhalation injury in this range of burn size independent of age, as demonstrated in .the three-dimensional probit analyses as compared to those patients with burn injury only. Furthermore, although most attention has been directed to mortality caused by inhalation injury in burn patients, the morbidity of inhalation injury and its effect on length of stay in hospital has not been realized totally. Through analysis of covariance, our data suggests that
Survival Without Inhalation Injury
£I 6.1
FIGS. 10 A and B. Three dimensional plot of probability of survival with inhalation injury (B).
following burn injury in patients without associated inhalation injury (A) and
726
TREDGET AND OTHERS
Length of Stay
50
401 (0
, 30
I~~~~~~~~~ "III--I I'-----I---"'I
co
cn
0
-j20 c
0
10
100
80 0
cU V
co 0
60
-J
X 40
Ann. Surg. December 1990
severe form of injury than detectable by bronchoscopy. In the presence of bronchoscopic pathology, this investigation added little to the diagnosis and, in their assessment, did not warrant the practical difficulties encountered in obtaining the results.5 In laboratory models of inhalation injury, Zawacki et al.2' and others22'23 found that severity of inhalation injury was a function of both the size of burn and the amount of smoke inhaled, and within limits carboxyhemoglobin levels were proportional to severity of injury. However the usefulness of carboxyhemoglobin levels as a predictor of the severity of inhalation injury has not been substantiated in burn patients,'8 in keeping with our own experience. Thus few clinical indices exist as yet that allow quantitative analysis of the magnitude of pulmonary injury after smoke inhalation. Early identification and quantification of the magnitude of injury in our inhalation-injured patients would allow identification of those patients who develop increased deadspace ventilation, prolonged requirements for mechanical ventilation, and are at risk of developing chronic pulmonary insufficiency, physiologic abnormalities that were documented in a subsegment of our inhalation-injured burn patients.24 As
20 77
78
79
80
81
82
83
84
85
86
Mean Length of Stay (With/Without Inhalation Injury)
87
Admission Year
FIG. 11. Length of stay in days versus year of admission. Top graph includes all admissions and the bottom graph shows the same population divided into those burn patients with and without inhalation injury. Data shown are the mean and the SEM for each year. The mean length of stay for the total number of patients was 26.2 ± 0.8 days. The mean length of stay for those patients without an associated inhalation injury was 23.7 ± 0.7 days compared to 74.4 ± 6.2 days for those patients with an inhalation injury (p < 0.01).
duration of stay in hospital is increased threefold on average for those surviving patients with inhalation injury and this prolongation of the morbid period is independent of the age (Fig. 13) and the size of the burn injury (Fig. 12). More importantly the duration of hospitalization in our center can be predicted within a very narrow range if patients have no respiratory tract involvement with their cutaneous burns, but when inhalation injury coexists, the mean length of stay is substantially longer and highly variable, as expressed by the wide standard error of the mean for each measurement (Figure 11). These factors become important in overall assessment of cost of care and make resource and manpower planning difficult.'9 They also illustrate the difficulty in assessing the severity ofthe respiratory tract 'burn' and the limitations of bronchoscopic assessment only in inhalation injury.20 Although Shirani et al.5 used 133Xenon lung scanning to detect respiratory tract damage in the absence of abnormalities at bronchoscopy, such cases represented a less
140
1201100
V)
80
0 c
co,
a)
60 |Inhalation Injury
40 20
0
1-14
30-69 TBSA (%)
15-29
70+
FIG. 12. Mean length of stay (LOS) versus %TBSA for small, moderate, large, and massive burns. Data shown are the mean for each group and the SEM. The mean LOS for those patients without an associated inhalation injury was 23.7 ± 0.7 days compared to 74.4 ± 6.2 days for those patients with an inhalation injury (p < 0.01).
Vol. 212 * No.6
ROLE OF INHALATION INJURY IN BURN TRAUMA
Mean Length of Stay (With/Without Inhalation Injury)
240 220
References
* -
200 180 .
160 .
3C10,
140
-
T
,
3J12080
Ko a -
Inhalation Injury ,
j
727
:
I
60
Age (years) FIG. 13. Mean length of stay (LOS) versus age for burn patients with and without inhalation injury. Data shown are the mean for each age grouping and the SEM. The mean LOS for those patients without an associated inhalation injury was 23.7 ± 0.7 days compared to 74.4 ± 6.2 days for those patients without an inhalation injury (p < 0.01).
the understanding of the pathophysiology of this injury increases, analysis of free radicals, prostanoids, or cytotoxic lymphokines in the pulmonary tissues'13'25 and the interaction of the burn wound with pulmonary function26 will allow increased understanding of both the severity of inhalation injury in burn patients and its treatment. Our data further documents that inhalation injury increases burn mortality independent of patient age and burn size. In surviving thermally injured adults and children, hospitalization is prolonged substantially by inhalation injury, again independent of the age and TBSA in a highly variable and as yet unpredictable fashion.
Acknowledgments The authors thank the residents, nurses, physiotherapists, occupational therapists, dieticians, staff physicians, and all others on the burn team for the excellent care they gave to the patients involved in this study.
1. Demling RH. Bums. N Engl J Med 1985; 313:1389-1393. 2. Traber EL, Linares HA, Hemdon DN. The pathophysiology of inhalation injury-a review. Bums 1988; 14:357-364. 3. Kinsella J. Smoke inhalation. Bums 1988; 14:269-279. 4. Gemmell CG, Pollok AJ, McMillan F, et al. Structural and functional changes in alveolar macrophages following the exposure of fire victims to smoke. Eur J Clin Invest 1987; 17:321-325. 5. Shirani KZ, Pruitt BA, Mason AD. The influence of inhalation injury and pneumonia on bum mortality. Ann Surg 1987; 20:82-87. 6. Baxter CR, Shires T. Physiologic response to crystalloid resuscitation of severe bums. Ann NY Acad Sci 1968; 150:874-880. 7. Hunt JL, Agee RN, Pruitt BA. Fiberoptic bronchoscopy in acute inhalation injury. J Trauma 1975; 15:641-649. 8. Thompson PB, Hemdon DN, Traber DL, et al. Effect on mortality of inhalation injury. J Trauma 1986; 26:163-165. 9. Margosches EH, Roi LD, Flora JD. The statistical analysis of bum patient data: a historical review. Bums 1978; 5:43-53. 10. Zawacki BE, Azen SP, Imbus SH, Chang YC. Multifactorial probit analysis of mortality in bumed patients. Ann Surg 1979; 189: 1 5. 11. Tompkins RG, Burke JF, Schoenfeld DA, et al. Prompt eschar excision: a treatment system contributing to reduced bum mortality. Ann Surg 1986; 194:272-281. 12. Merrell SW, Saffle JR, Sullivan JJ, et al. Increased survival after major thermal injury. Am J Surg 1987; 154:623-627. 13. Tompkins RG, Remensnyder JP, Burke JF, et al. Significant reductions in mortality for children with bum injuries through the use of prompt eschar excision. Ann Surg 1988; 208:33-41. 14. Stem M, Waisbren BA. A method by which bum units may compare their results with a base line curve. Surg Gynecol Obstet 1976; 142:230-234. 15. Flora JD. A method for comparing survival of bum patients to a standard survival curve. J Trauma 1976; 18:701-705. 16. Iglehart JK. Canada's health care system. N Engl J Med 1986; 315: 202-208. 17. Cox DR. Analysis of Binary Data. London: Chapman and Hall, 1970, pp 14-29. 18. Clark WR, Bonaventura M, Myers W. Smoke inhalation and airway management at a regional bum unit: 1974-1983. J Bum Care Rehab 1989; 10:52-62. 19. Curreri PW, Luterman A, Braun D, Shires GT. Bum injury: analysis of survival and hospitalization time for 937 patients. Ann Surg 1980; 192:472-478. 20. Bingham HG, Gallagher TJ, Powell MD. Early bronchoscopy as a predictor of ventilatory support for bumed patients. J Trauma 1987; 27:1286-1289. 21. Zawacki BE, Jung RC, Joyce J, Rincon E. Smoke, bums and the natural history of inhalation injury in fire victims: a correlation ofexperimental and clinical data. Ann Surg 1977; 185:100-110. 22. Hemdon DN, Traber DL, Niehaus GD, et al. The pathophysiology of smoke inhalation injury in a sheep model. J Trauma 1984; 24:1044-1051. 23. Shimazu T, Yukioka T, Hubbard GB, et al. A dose-responsive model of smoke inhalation injury. Ann Surg 1987; 206:89-97. 24. Stollery DE, Jones RL, King EG. Deadspace ventilation: a significant factor in respiratory failure after thermal inhalation. Crit Care Med 1987; 15:260-261. 25. Clark CJ, Pollock AJ, Reid WH, et al. Role of pulmonary alveolar macrophage activation in acute lung injury after bums and smoke inhalation. Lancet 1988; ii:872-877. 26. Demling RH, Will JA, Belzer FO. Effect of major thermal injury on the pulmonary microcirculation. Surgery 1978; 83:746-751.
