Frostbite injuries: independent predictors of outcomes
Morgan Schellenberg, Vincent Cheng, Kenji Inaba, Christopher Foran, Zachary Warriner, Marc D. Trust, Damon Clark, Demetrios Demetriades
Department of Trauma and Acute Care Surgery, LAC+USC Medical Center, Los Angeles, United States of America
Objective: Frostbite injuries are important causes of morbidity and mortality after trauma. Epidemiology, injury patterns, and outcomes after frostbite among patients presenting to trauma centers are incompletely defined. The purpose of this study was to delineate patient demographics, clinical characteristics, and independent predictors of outcomes after frostbite.
Material and Methods: Patients with frostbite injury were identified from the National Trauma Data Bank (NTDB) (2007-2014). Demographics, clinical/injury data, and outcomes were collected. Patients were dichotomized into study groups based on intensive care unit (ICU) admission. Univariate analysis was performed with the Mann-Whitney U, Fisher’s exact, or Chi-Square test as appropriate. Multivariate analysis using logistic regression determined independent predictors of outcomes.
Results: Over the study period, 241 patients were identified. Median body temperature on admission was 36.3⁰C (IQR 33.4-36.7). Mortality was 3% (n= 7). ICU admission was required in 101 (42%) patients and 48 (20%) underwent surgical intervention. On multivariate analyses, mortality was predicted by lower admission GCS (p= 0.027) and amputation by higher HR (p= 0.013). Need for ICU admission was predicted by older age (p= 0.010), male gender (p= 0.040), higher HR (p= 0.031) and ISS (p< 0.001), and lower GCS (p= 0.001). Prolonged hospital LOS was predicted by higher heart rate (p< 0.001) and ISS (p< 0.001).
Conclusion: Frostbite injuries are uncommon but can necessitate surgical intervention and cause mortality. Lower GCS and higher heart rate, but not body temperature, portend poor outcomes. These findings can be used to triage patients appropriately upon admission and to better inform prognosis after frostbite injuries.
Keywords: Frostbite, thermal injuries, hypothermia
Frostbite injuries are infrequent but important causes of morbidity and mortality after trauma. In part because of their rarity, frostbite injuries are incompletely defined by the current literature (1). Knowledge of predictors of outcomes, such as mortality and need for amputation, would be useful for prognostication and to anticipate resource utilization after this uncommon mechanism of injury.
The primary objective of this study was to define independent predictors of outcomes [mortality, need for surgical intervention, need for intensive care unit (ICU) admission, ICU length of stay (LOS), and hospital LOS] after frostbite injuries using a large, nationwide patient population from the National Trauma Data Bank (NTDB) (2). The secondary objectives were to define the epidemiology and clinical characteristics of patients who sustain frostbite injuries. Our hypothesis was that worsened physiologic status on presentation to hospital [e.g. lower body temperature, tachycardia, hypotension, diminished glasgow coma scale (GCS) score] would be predictive of poor outcomes after frostbite.
Material and Methods
In this retrospective observational study, all patients with frostbite injuries (AIS codes 915000, 915002, 915004, 915006) between January 1, 2007 and December 31, 2014 were identified from the NTDB. The NTDB, run by the American College of Surgeons (ACS), is the largest trauma registry in existence and is comprised of aggregate data from centers across the United States (2). There were no exclusion criteria. Ethics committee approval by the Institutional Review Board of the University of Southern California was obtained. Informed consent was waived due to the study’s retrospective observational design of deidentified patient data.
After patient identification, demographics (age, gender), clinical data [admission heart rate (HR), systolic blood pressure (SBP), temperature, and GCS score], and injury data [Injury Severity Score (ISS)] were collected from the NTDB. Primary outcome was mortality. Secondary outcomes included need for surgical intervention (debridement, amputation), need for ICU admission, ICU LOS, and hospital LOS.
Patients were dichotomized into study groups based on the need for ICU admission or not. Continuous variables are represented as median [interquartile range (IQR)] and categorical variables are given as number (percentage). Univariate analysis was used to compare patient demographics, clinical/injury characteristics, and outcomes between study groups using the Mann-Whitney U test, Fisher’s exact test, or Chi-square test as appropriate. Multivariate regression with the enter method included all clinically relevant predictor variables. Data were collected and analyzed using SPSS version 20.0 (IBM Corporation, Armonk, NY). Statistical significance was defined as p< 0.05.
Patient Demographics, Clinical Data, Injury Data, and Outcomes
Overall, 241 (< 0.01%) patients with frostbite injury were identified from the NTDB (Figure 1). ICU admission was required in 101 (42%) patients. Median age was 44 years [interquartile range (IQR) 29-58] and 184 (76%) patients were male (Table 1). Median temperature on arrival to the emergency department was 36.3°C (IQR 33.4-36.7). In general, patients were hemodynamically normal and intact neurologically. Median ISS was 8 (IQR 2-14). Overall mortality was 3% (n= 7) (Table 2). Surgical intervention was required in 20% of the patients (n= 48). Overall median hospital LOS was 5 days (IQR 2-10).
|total patients (n= 241)||ICu admission (n= 101, 42%)||no ICu admission (n= 140, 58%)||p|
|Age (years)||44 (29-58)||48 (37-63)||41 (28-55)||0.008|
|Gender (male)||184 (76%)||81 (80%)||103 (74%)||0.283|
|HR (bpm)||96 (84-108)||95 (83-108)||96 (85-95)||0.546|
|SBP (mmHg)||133 (117-149)||130 (111-149)||134 (122-149)||0.165|
|Temperature (0C)||36.3 (33.4-36.7)||34.1 (30.6-36.3)||36.6 (34.1-35.9)||< 0.001|
|GCS||15 (13-15)||13 (6-15)||15 (15-15)||< 0.001|
|ISS||8 (2-14)||14 (7-21)||5 (1-9)||< 0.001|
|ICU: Intensive care unit; patient was admitted to ICU after presentation to the emergency department, HR. Heart rate, Bpm: Beats per minute, SBP: Systolic blood pressure, GCS: Glasgow Coma Scale score, ISS: Injury Severity Score. Continuous variables expressed as median [interquartile range]; p-value calculated with Mann-Whitney U-test. Categorical variables expressed as number (%); p-value calculated with Fisher’s Exact Test or Pearson Chi Square Test as appropriate.|
|total patients (n= 241)||ICu admission (n= 101, 42%)||no ICu admission (n= 140, 58%)||p|
|Mortality||7 (3%)||7 (7%)||0 (0%)||0.004|
|Need for surgical Intervention||48 (20%)||23 (23%)||25 (18%)||0.414|
|Debridement||43 (18%)||19 (19%)||24 (17%)||0.737|
|Amputation||13 (5%)||9 (9%)||4 (3%)||0.047|
|Upper extremity||4 (2%)||3 (3%)||1 (1%)||0.312|
|Lower extremity||11 (5%)||8 (8%)||3 (2%)||0.056|
|Hospital LOS||5 (2-10)||8 (4-16)||3 (2-6)||< 0.001|
|ICU LOS||0 (0-3)||3 (2-7)||-||-|
|ICU: Intensive care unit; patient was admitted to ICU after presentation to the emergency department, LOS: Length of stay in days. Continuous variables expressed as median [interquartile range]; p-value calculated with Mann-Whitney U-test. Categorical variables expressed as number (%); p-value calculated with Fisher’s Exact Test or Pearson Chi Square Test as appropriate.|
Patients Who Required ICU Admission vs. Those Who Did Not
Patients who were admitted to the ICU (n= 101, 42%) tended to be older (48 vs. 41 years, p= 0.008), have a lower admission body temperature (34.1 vs. 36.6°C, p< 0.001), and a lower admission GCS (13 vs. 15, p< 0.001) than patients who were not admitted to the ICU (n= 140, 58%) (Table 1). Patients admitted to the ICU had a higher mortality (7% vs. 0%, p= 0.004), greater need for amputation (9% vs. 3%, p= 0.047), and longer hospital stay (8 vs. 3 days, p< 0.001) than patients not admitted to the ICU (Table 2).
Independent Predictors of Outcomes
Mortality: On multivariate logistic regression, only lower admission GCS was independently predictive of mortality (p= 0.027) (Table 3).
|Gender (male)||< 0.001||< 0.001-> 999.999||0.997|
|OR: Odds ratio, CI: Confidence interval, ICU: Intensive care unit; patient was admitted to ICU after presentation to the emergency department. HR: Heart rate in beats per minute. SBP: Systolic blood pressure in mmHg, GCS: Glasgow Coma Scale score, ISS: Injury Severity Score, LOS: Length of stay in days. Multivariate analysis was performed using logistic regression.|
Need for surgical intervention: No independent predictors of the need for debridement were identified. Higher admission HR was independently predictive of the need for amputation (p= 0.013).
ICU admission and LOS: The need for ICU admission was predicted by older age (p= 0.010), male gender (p= 0.040), higher admission HR (p= 0.031) and ISS (p< 0.001), and lower admission GCS (p= 0.001). Longer ICU LOS was predicted by older age (p< 0.001); male gender (p= 0.030); higher HR (p< 0.001) and ISS (p< 0.001); and lower SBP (p= 0.003) and GCS (p< 0.001) (Table 4).
|Age (years)||0.010||0.004-0.015||< 0.001|
|SBP||-0.005||-0.009 - -0.002||0.003|
|GCS||-0.078||-0.111 - -0.045||< 0.001|
|RC: Regression coefficient, CI: Confidence interval, ICU: Intensive care unit; patient was admitted to ICU after presentation to the emergency department. HR: Heart rate in beats per minute, SBP: Systolic blood pressure in mmHg, GCS: Glasgow Coma Scale score, ISS: Injury Severity Score, LOS: Length of stay in days. Multivariate analysis was performed using logistic regression.|
Hospital LOS: Higher admission HR (p< 0.001) and ISS (p< 0.001) were independent predictors of prolonged hospital LOS.
Frostbite, wherein skin is exposed to freezing temperatures and injured by cold thermal energy, is relatively common in some parts of the world, with a lifetime incidence in colder climates of approximately 10% (1,3,4). Most of the available data on frostbite injuries originate from small case series from the military or from adventurists (5-10). The generalizability of these studies, particularly in terms of their applicability to the general population, remains unclear. Existing literature is scarce on patient epidemiology and severity of illness after frostbite injury among civilians. Additionally, independent predictors of outcomes, such as the need for surgical intervention, ICU admission, and mortality, are undefined.
This study examined a nationwide patient population sustaining frostbite injuries and found that frostbite is an infrequent cause of presentation to American hospitals. Patients who suffered from these injuries tended to be middle aged men. Despite exposure to freezing temperatures, most were normothermic upon presentation to the Emergency Department. Mortality in this population was low but surgical intervention, either with debridement or amputation, was required in 20% of the patients. These data would suggest that systemic sequelae after frostbite injuries are infrequent but that local complications such as tissue necrosis and infection are relatively common.
Elevated heart rate and lower GCS on presentation to hospital were almost uniformly associated with poorer outcomes, including mortality, need for surgical intervention, need for ICU admission, and prolonged hospital stay. Other risk factors for worse outcomes identified by the present study were older age, male gender, lower SBP on admission, and higher ISS. Interestingly, body temperature on admission was not predictive of any study outcome. Taken together, these findings suggest that many of the clinical parameters known to portend worse outcomes among trauma patients, such as tachycardia, hypotension, depressed neurological status, and older age, also herald worse outcomes among the subset of trauma patients with frostbite injuries.
Limitations of the present study must be acknowledged. Retrospective studies are inherently limited by their design. Additionally, the NTDB lacks sufficient granularity to identify the depth or severity of the frostbite injury, information which may be useful in prognostication after injury (11). Finally, nonsurgical treatment strategies, such as the type of rewarming employed or the use of thrombolytics, are not coded by the NTDB. Therefore, the effect of management strategy on outcomes after frostbite is not captured by this study.
Frostbite injuries are infrequent among the American civilian population. Despite having a very low mortality rate, approximately half of the patients with frostbite will require ICU admission and one fifth will require surgical intervention. Although a number of independent predictors of poor outcomes were identified, higher heart rate and depressed GCS on admission in particular portended worse outcomes. Clinicians should be especially vigilant with patients who arrive to the ED with these clinical features after frostbite injury. The knowledge of patient epidemiology, clinical characteristics, and predictors of outcomes imparted by this study may be useful in planning resource utilization and to better inform conversations with patients and families about prognosis after frostbite injury.
Cite this article as: Schellenberg M, Cheng V, Inaba K, Foran C, Warriner Z, Trust MD, et al. Frostbite injuries: independent predictors of outcomes. Turk J Surg 2020; 36 (2): 218-223.
Ethics committee approval by the Institutional Review Board of the University of Southern California was obtained (Approval granted January 28, 2019; Approval number is: HS-19-00015).
Informed consent was waived due to the study’s retrospective observational design of deidentified patient data.
Concept – M.S., V.C.; Design – M.S., V.C., C.F., Z.W., M.D.T.; Supervision – M.S.; Resource – M.S., K.I., D.C., D.D.; Materials - M.S., K.I., D.C., D.D.; Data Collection and/or Processing - M.S., V.C., C.F., Z.W., M.D.T.; Analysis and/or Interpretation - M.S., V.C.; Literature Search M.S.; Writing Manuscript – M.S., V.C., K.I.; Critical Reviews – C.F., Z.W., M.D.T., D.C., D.D.
The authors have no conflicts of interest to declare.
The authors declared that this study has received no financial support.
- Nygaard RM, Endorf FW. Frostbite in the United States: an examination of the national burn repository and national trauma data bank. J Burn Care Res 2018; 39: 780-5.
- American College of Surgeons. National trauma data bank. 2019. Available at https://www. ntdbdatacenter.com/ [accessed 7 Jan 2019].
- Murphy JV, Banwell PE, Roberts AHN, McGrouthier DA. Frostbite: pathogenesis and treatment. J Trauma 2000; 48 (1): 171-8.
- Makinen TM, Jokelainen J, Nayha S, Laatikainen T, Jousilahti P, Hassi J. Occurrence of frostbite in the general population – work-related and individual factors. Scand J Work Environ Health 2009; 35: 384-93.
- Heil KM, Oakley EH, Wood AM. British military freezing cold injuries: a 13-year review. J R Army Med Corps 2016; 162: 413-8.
- Moran DS, Heled Y, Shani Y, Epstein Y. Hypothermia and local cold injuries in combat and non-combat situations – The Israeli experience. Aviat Space Environ Med 2003; 74 (3): 281-4.
- Lorentzen AK, Penninga L. Frostbite – A case series from arctic Greenland. Wilderness Environ Med 2018; 29 (3): 392-400.
- Pandey P, Vadlamudi R, Pradhan R, Pandey KR, Kumar A, Hackett P. Case report: severe frostbite in extreme altitude climbers – the Kathmandu Iloprost experience. Wilderness Environ Med 2018; 29 (3): 366-74.
- Strohle M, Rauch S, Lastei P, Brodmann Maeder M, Brugger H, Paal P. Frostbite injuries in the Austrian Alps: a retrospective 11-year national eegistry study. High Alt Med Biol 2018; 19 (4): 316-20.
- Harirchi I, Arvin A, Vash JH, Zafarmand V. Frostbite: incidence and predisposing factors in Mountaineers. Br J Sports Med 2005; 39: 898-901.
- Handford C, Thomas O, Imray CHE. Frostbite. Emerg Med Clin North Am 2017; 35: 281-99.