Article, Surgery

The utility of abnormal initial arterial blood gas values in determining clinical futility of trauma cases with severe hemorrhage

a b s t r a c t

Background: Patients who experience trauma with severe hemorrhage requiring immediate surgery and massive blood transfusion often present with markedly abnormal laboratory values. These cases require valuable re- sources; however, little is known regarding prognostic factors that correlate with mortality. The purpose of this study was to determine whether abnormal initial arterial blood gas pH, a marker for severe blood loss, could serve as a prognostic indicator for these patients.

Methods: An IRB approved retrospective study was performed at LAC + USC Medical Center Level I Trauma Cen- ter. Data was collected from trauma patients with severe hemorrhage admitted between June 2015 and April 2016 who were immediately admitted to the OR following entry into the ER. Baseline variables of age, sex and mechanism of trauma were collected. The pH readings from the initial three ABG data were obtained, and mor- tality was determined for each patient.

Results: We identified 247 patients, 84.2% of which were male. Ages ranged from 1 to 91 years (average = 38.4). Overall mortality was 13.8%. The average Initial pH value for non-survivors (7.10 +- 0.13) was significantly lower than for survivors (7.34 +- 0.07) [p b 0.001]. Among patients whose initial three ABG pH values averaged <=7.15, the survival rate was 8.7%. Ten patients had any single recorded pH value <= 6.91. The mortality rate among these patients was 90%.

Conclusions: Consideration should be given to initial pH values when resuscitating “red blanket” patients. How- ever, the pH values alone cannot reliably be used to determine clinical futility in individual patients in the early period after injury.

(C) 2018


Hemorrhage is the leading cause of Preventable death in civilian trauma centers [1,2]. Improvements in prehospital care, resuscitation and operative management of trauma patients have improved survival rates of severely injured patients [3]. massive transfusion and the utili- zation of other valuable Hospital resources are being increasingly used [3]. A particular group of patients requiring intensive care and resource allocation are those that are admitted to the operating room immediate- ly following entry to the Emergency Department. These are termed “red blanket” patients at some hospitals, including the LAC + USC Medical Center Level I Trauma Center [4-6]. These cases involve the mobilization of specialist teams, massive transfusion protocols, and operating the- atres for the imminent arrival of a critically ill patient. In most cases, these patients are admitted after major trauma with subsequent blood loss. In relation to the severity of the injuries sustained and the amount

* Corresponding author at: 1520 San Pablo Street, Suite 3451, Los Angeles, CA 90033, USA.

E-mail address: [email protected] (J.M. Berger).

of hemorrhage, such patients have low and/or declining arterial blood gas pH levels. The adverse effects of prolonged acidosis are wide- ranging and far-reaching within the human system. cerebral function is closely tied to metabolism, such that traumatic brain injury disturbs acid-base homeostasis and acid-base derangements themselves have a deleterious effect on neurologic status [7]. Thus, arterial blood gas values have a potential indication regarding not only mortality but long-term patient outcomes.

Many resources are spent on “red blanket” cases admitted following major trauma. However, little data exists to validate its effectiveness on patients in whom the severity of injury and blood loss (directly correlat- ing with detrimentally low pH levels) is past a certain, undetermined cutoff. Identification of new prognostic factors in these cases could aid in clinical decision-making for Healthcare workers in all facets of healthcare delivery in the ER to the OR. A theoretical “cut-off” point of abnormal laboratory values could allow physicians to reliably terminate futile care, conserving blood products and other resources for patients with increased Probability of survival. Additionally, predictive laborato- ry values could help with planning the long-term care of surviving pa- tients who are affected by the long-term sequelae of massive blood loss and acidosis.

0735-6757/(C) 2018

1254 A. Katirai et al. / American Journal of Emergency Medicine 36 (2018) 12531256

The purpose of this study was to determine whether initial arterial blood gas pH levels correlate with mortality in cases of trauma associat- ed with severe hemorrhage that require massive transfusion. Our retro- spective study examined “red blanket” patients presenting to Los Angeles county hospital, a Level I trauma center. We collected data re- garding patient demographics, mechanism of trauma, initial Arterial blood gas values, and operation performed. Our results indicated that initial ABG pH readings significantly differed between cases of sur- vival and cases of mortality. To our knowledge, this is the first study of “red blanket” patients regarding their ABG readings as a potential prog- nostic indicator.


We conducted an IRB-approved, retrospective study. No interventions were performed as a part of this study. All patient information was de- identified prior to data gathering and analysis. We identified 682 patients admitted to the Los Angeles County Hospital Acute Care Surgery operat- ing room directly from the Emergency Department between June 1, 2015 and April 1, 2016. Patients admitted whose diagnoses were un- known or did not include trauma, patients whose trauma was not associ- ated with severe hemorrhage, and patients whose initial ABG values could not be determined were excluded. This resulted in 247 patients re- maining for analysis, with an average ABG pH of 7.30 +- 0.11 in the entire group. Using a two-sample, one-tailed analysis for statistical power, this sample size was determined to be sufficient to detect an effect on mortal- ity of pH <= 7.15 (sample size of 59 needed for statistical power N 95%).

Electronic medical records of included patients were retrospectively analyzed. Information gathered for each patient included age, sex, mech- anism of trauma, type of procedure performed in the OR, and the first three ABG pH levels within the first 24 h of admission recorded. The type of procedure performed in the OR was defined by the primary proce- dure code entered in the patient’s electronic medical record. We included patients with only one or two available initial ABG pH readings. We also determined mortality prior to discharge from the hospital. Microsoft Excel (Microsoft Corporation, Redmond, WA) was utilized to calculate av- erages and standard deviations. Student’s two-tailed t-test was used to determine p values. MedCalc (MedCalc Software, Ostend, Belgium) was used to construct the receiver operating characteristic (ROC) curve. The electronic medical record search and data analysis was completed by two authors (A.K., M.J.L.) and reviewed by the senior author (J.M.B.).


A total of 247 patients met inclusion criteria and were included in retrospective analysis (Fig. 1). Of the 45 patients who were excluded

Fig. 1. Flowchart of patient selection for final analysis. Initial search included all patients admitted to the Los Angeles County Hospital Acute Care Surgery operating room directly from the Emergency Department between June 1, 2015 and April 1, 2016.

Fig. 2. (A) Frequency of primary diagnoses listed on electronic medical record of “red blanket” case. (B) Frequency of primary procedures coded for in electronic medical record of “red blanket” case.

from the study due to an absence of available ABG results, the mortality rate was 26.7%. Of the included patients, 208 patients were male (84.2%) and 39 were female (15.8%). The average age was 38.4 years (range 1-91 years, standard deviation +- 16.9 years).

The most common mechanisms of trauma were gunshot wound(s) (28%), Stab wound(s) (24%), motor vehicle accident (16%), and auto vs. pedestrian accident (11%) (Fig. 2A). The most common pri- mary procedure performed in the operating room was an Exploratory laparotomy (69%), followed by exploration of traumatic wound (9%) and exploration of neck artery (5%) (Fig. 2B).

The mortality rate among patients included in this study was 13.8% (Table 1). The average age of survivors was 37.4 +- 15.8 years and the average age of nonsurvivors was 46.3 +- 21.1 years. The average of the first three arterial blood gas pH results was found to be significantly lower in cases of nonsurvival compared to cases of survival (7.10 +- 0.13 vs. 7.34 +- 0.07, p b 0.001). The average ABG pH values were consis- tently lower in nonsurvivors compared to survivors in the first, second, and third ABG results (Fig. 3). In both survivors and nonsurvivors, the

Table 1

Comparison of data between survivors and nonsurvivors of trauma with severe hemor- rhage requiring direct transfer to the operating room from the Emergency Department. Patients were considered a mortality if they were declared deceased prior to discharge from the hospital for their Traumatic event. Average pH values were acquired from the first three arterial blood gas readings.

Survivors (n = 213)

Nonsurvivors (n = 34)

Average age

37.4 +- 15.8

46.3 +- 21.1

Percent male

Average arterial pH



7.34 +- 0.07


7.10 +- 0.13

* p b 0.001.

Katirai et al. / American Journal of Emergency Medicine 36 (2018) 12531256 1255

Fig. 3. Averages of the first (pH 1), second (pH 2), and third (pH 3) ABG pH values. Error bars represent standard error of the mean. Blue dashed line = survivors, red solid line = nonsurvivors.

average second and third arterial blood gas pH were higher than the av- erage initial ABG pH.

The distribution of average ABG pH values among nonsurvivors was also wider than the distribution of ABG pH values among survivors (Fig. 4). Additionally, while 21 nonsurvivors (61.8% of nonsurvivors) had an average ABG pH of 7.15 or less, only two survivors (0.9% of sur- vivors) had an average ABG pH in that range. Therefore, among patients whose initial three ABG pH values averaged 7.15 or less, the survival rate was only 8.7%. When using an average pH of 7.15 as a cut-off to de- termine eventual mortality, the sensitivity of the screening test was 61.8% and the specificity was 99.1%. We constructed an ROC curve using average ABG pH values (Fig. 5). The area under the curve was 0.958 (95% Confidence Interval 0.925 to 0.979, p b 0.0001).

No survivors had an average ABG pH of 7.00 or less, while nine nonsurvivors (26.5% of nonsurvivors) fell into this range. On the other hand, all patients with an average ABG pH of 7.36 or higher survived. Ten patients had any single recorded pH value <= 6.91. The mortality rate among these patients was 90%.


Profound trauma causing hemorrhage and hemodynamic instability is an important indication for the initiation of a “red blanket” protocol, involving the transfer of a patient from the emergency room directly to the operating room and the initiation of massive transfusion. Injured patients requiring massive transfusion often initially present with ab- normal physiologic and laboratory data. Analyzing the causes, presenta- tions, and outcomes of these cases can aid physicians in perioperative planning and in predicting mortality. Defining medical futility in these

scenarios could be important in resource-limited settings and with regards to preventing the excess expenditure of time and materials. This will allow physicians to determine whether initial laboratory data is compatible with eventual survival when making withdrawal-of-Care decisions in conjunction with patient family members [3].

Previously identified prognostic criteria associated with futility in cases of trauma are mechanism of injury, age, gender, comorbidities, ad- mission SBP, core temperature, lactic acid levels, and Base deficit [8-11]. The likelihood of futile resuscitation efforts in association with easily identifiable clinical criteria at admission has been investigated in burn patients and in patients requiring thoracotomies [10,12]. However, little information exists regarding criteria to define the futility of “red blan- ket” protocol initiation. Arterial blood gas results are commonly ac- quired in the emergent setting of a “red blanket” case, but their role in predicting outcomes is yet to be elucidated. Examining the correlation of ABG results and outcomes could help physicians define medical futil- ity. Along with helping to predict mortality, preoperative laboratory values could aid physicians in determining potential postoperative mor- bidity, such as Neurological deficits attributable to long periods of hyp- oxia and acidosis. An additional consideration of defining futility might also be given to the probability of successful Organ donation. One study found that 37.7% of brain dead donors had metabolic acidosis at the Initial diagnosis of brain death, and that the presence of metabolic acidosis correlated with a decreased number of organs retrieved from a patient [13].

Acidosis in cases of hypovolemic shock caused by severe hemor- rhage reflects metabolic derangements in cells experiencing inadequate oxygenation. tissue hypoperfusion leads to the generation of large quantities of hydrogen ions from lactic acid and other unmeasured an- ions, such as TCA cycle intermediates, resulting in metabolic acidosis [14,15]. Lactic acidosis in the brain plays an important role in the devel- opment of cerebral edema and gross neuronal damage [16]. Patients in hemorrhagic shock are vulnerable to irreversible neuronal cell damage dependent on the duration and severity of lactic acidosis and ischemia [17]. Thus, the degree of acidosis might correlate not only with survival but with the future rate of Neurologic recovery or the extent of irrevers- ible brain damage following hemorrhagic shock. Additionally, acidosis has been postulated to contribute to cardiac arrhythmia and increase Pulmonary vascular resistance [18,19]. Severe and prolonged lactic aci- dosis in conjunction with global hypoxia has a broad range of adverse effects on the human body. Additional clinical and laboratory studies should be undertaken to better understand the mechanism of organ in- jury and recovery in cases of hypovolemic shock to better predict out- comes and tailor treatment following resuscitation.

In our group, the most common mechanisms of injury that lead to the initiation of the “red blanket” protocol were gunshot wounds, Stab wounds, and motor vehicle accidents. The most common primary pro- cedure performed was an exploratory laparotomy. Our results indicate that initial ABG pH levels significantly differ between cases of survival

Fig. 4. Distribution of patients based on the average of the initial three ABG pH values. Percentages displayed represents percent of each group (nonsurvivors or survivors) that had an average pH value within the given range.

1256 A. Katirai et al. / American Journal of Emergency Medicine 36 (2018) 12531256

procedure performed, as multiple procedures may have been per- formed under a single primary procedure coded for in the electronic medical record. functional outcomes in survivors were not assessed. Confounding factors such as the previous health of the patient, patient age, or simultaneous traumas not associated with hemorrhage (such as head trauma) were also not accounted for. Future studies should stratify initial laboratory results based on patient demographic informa- tion, medical history, mechanism of trauma, and surgical complexity.

Despite these limitations, our study provides an initial analysis of “red blanket” cases at the LAC + USC Medical Level I Trauma Center. Though this study represents only a preliminary analysis of “red blan- ket” trauma patients with severe hemorrhage, it suggests that future studies may be able to identify ABG pH values as a predictive factor of mortality.

Fig. 5. Receiver operating characteristic (ROC) curve. Curve was constructed using average of initial three ABG pH values, with mortality defined as a positive outcome. (AUC = area under the curve).

and cases of nonsurvival in “red blanket” patients who have experi- enced trauma associated with severe hemorrhage. Average ABG pH levels in nonsurvivors were consistently lower in all three initial read- ings compared to that of survivors. In our limited group of cases, an av- erage ABG pH of 7.00 or lower correlated with a 100% mortality rate. However, several cases of survival with profoundly decreased arterial pH have been reported [3,18]. Thus, our data alone cannot confidently indicate a “cut-off” point that defines medical futility. However, our sta- tistical analysis indicates that ABG pH could perform well as an informa- tive diagnostic test for mortality [20]. The high specificity associated with using an average ABG pH of 7.15 as a cut-off value for screening suggests that an average ABG pH above this level could be a useful pos- itive prognostic indicator. The results of this study may validate the use of arterial blood gas measurements pre-hospital resuscitation efforts, providing a tool for early triage, guiding the selection of fluid resuscita- tion, and delivering useful information to physicians regarding a patient’s status prior to activation of a “red blanket” protocol [21]. Using arterial blood gas parameters to help direct prehospital resuscita- tion could help predict and prevent multiple organ failure, irreversible brain damage, or mortality.

This study was subject to several limitations. The mortality rate of patients for whom no ABG value was acquired (26.7%) was higher than the mortality rate of patients included in this study, and arterial blood gas values may have had a different distribution among these pa- tients. The sample size of nonsurvivors in this study was relatively small

(34). ABG sample acquisition and sample processing may have differed between patients. Additionally, our study did not account for the time of the initial ABG acquisition relative to the time of the traumatic event nor did it account for the gap in time between subsequent ABG results. Our study also did not take into consideration the complexity of the surgical


  1. Borgman MA, Spinella PC, Perkins JG, et al. The ratio of blood products transfused af- fects mortality in patients receiving massive transfusions at a combat support hospi- tal. J Trauma 2007;63(4):805-13.
  2. Acosta JA, Yang JC, Winchell RJ, et al. Lethal injuries and time to death in a level I trauma center. J Am Coll Surg 1998;186(5):528-33.
  3. Barbosa RR, Rowell SE, Diggs BS, et al. Profoundly abnormal initial physiologic and biochemical data cannot be used to determine futility in massively transfused trau- ma patients. J Trauma 2011;71(2 Suppl. 3):S364-369.
  4. Snoek S, Butson B, Wittenberg M. A challenging penetrating trauma case. Air Med J 2016;35(2):88-92.
  5. Winearls J, Wullschleger M, Wake E, et al. Fibrinogen early in severe trauma studY (FEISTY): study protocol for a randomised controlled trial. Trials 2017;18(1):241.
  6. Roemer MI, Moustafa AT, Hopkins CE. A proposed hospital quality index: hospital Death rates adjusted for case severity. Health Serv Res 1968;3(2):96-118.
  7. Clausen T, Khaldi A, Zauner A, et al. Cerebral acid-base homeostasis after severe trau- matic brain injury. J Neurosurg 2005;103(4):597-607.
  8. Nirula R, Gentilello LM. Futility of resuscitation criteria for the “young” old and the “old” old trauma patient: a National Trauma Data Bank analysis. J Trauma 2004; 57(1):37-41.
  9. Cothren CC, Moore EE. Emergency department thoracotomy for the critically injured patient: objectives, indications, and outcomes. World J Emerg Surg 2006;1:4.
  10. Krishna G, Sleigh JW, Rahman H. Physiological predictors of death in exsanguinating trauma patients undergoing conventional trauma surgery. Aust N Z J Surg 1998; 68(12):826-9.
  11. Cerovic O, Golubovic V, Spec-Marn A, Kremzar B, Vidmar G. Relationship between injury severity and lactate levels in severely injured patients. Intensive Care Med 2003;29(8):1300-5.
  12. Fratianne RB, Brandt CP. Determining when care for burns is futile. J Burn Care Rehabil 1997;18(3):262-7 [discussion 260-261].
  13. Lee JH, Kim MS, Na S, Koh SO, Sim J, Choi YS. Evaluation of acid-base status in brain dead donors and the impact of metabolic acidosis on organ retrieval. Minerva Anestesiol 2013;79(9):1011-20.
  14. Hobbs TR, O’Malley JP, Khouangsathiene S, Dubay CJ. Comparison of lactate, base ex- cess, bicarbonate, and pH as predictors of mortality after severe trauma in rhesus macaques (Macaca mulatta). Comp Med 2010;60(3):233-9.
  15. Bruegger D, Kemming GI, Jacob M, et al. Causes of metabolic acidosis in canine hem- orrhagic shock: role of unmeasured ions. Crit Care 2007;11(6):R130.
  16. Staub F, Mackert B, Kempski O, Peters J, Baethmann A. Swelling and death of neuro- nal cells by lactic acid. J Neurol Sci 1993;119(1):79-84.
  17. Rehncrona S, Rosen I, Siesjo BK. Brain lactic acidosis and ischemic cell damage: 1. Biochemistry and neurophysiology. J Cereb Blood Flow Metab 1981;1(3):297-311.
  18. Opdahl H. Survival put to the acid test: extreme arterial blood acidosis (pH 6.33) after Near drowning. Crit Care Med 1997;25(8):1431-6.
  19. Orchard CH, Cingolani HE. Acidosis and arrhythmias in cardiac muscle. Cardiovasc Res 1994;28(9):1312-9.
  20. Safari S, Baratloo A, Elfi M, Negida A. Evidence based emergency medicine; part 5 re- ceiver operating curve and area under the curve. Emerg (Tehran) Spring 2016;4(2): 111-3.
  21. Jousi M, Reitala J, Lund V, Katila A, Leppaniemi A. The role of pre-hospital blood gas analysis in trauma resuscitation. World J Emerg Surg 2010;5:10.