Efficient resource management presents an ongoing challenge for trauma centers. This process could be facilitated by a readily available marker of potential resource consumption. This indicator should identify patients more likely to require transfusion (blood bank resources) and intensive care services, to develop complications, and to require extended hospital stays. Such an indicator should be more precise than other admission data (vital signs, anatomic injury) and thus could potentially be used to assist in early identification of these critically injured patients.
Base deficit (BD) has been shown to be a valuable indicator of shock and the efficacy of resuscitation.  The severity of admission BD correlates with the volume of fluid required for resuscitation  and the presence of intra-abdominal hemorrhage. [3,4] BD has also been demonstrated to be predictive of mortality after trauma. [5,6]
This study was performed to determine whether admission BD was indicative of resource utilization in the care of patients at a trauma center. Specifically, could admission BD be used to identify those patients likely to require blood transfusion in the first 24 hours, and therefore help to guide ordering practices for laboratory tests (type and hold vs. type and cross-match)? Additionally, is BD indicative of anatomic injury severity, physiologic status, and predicted probability of survival? Furthermore, can BD be used to identify patients at high risk for shock-related complications such as adult respiratory distress syndrome (ARDS), coagulopathy, renal failure, and/or multiorgan system failure (MOF)?
Patients admitted to the trauma service at UCSF/Fresno Valley Medical Center who had an arterial blood gas obtained within 1 hour of admission were included. Patients under age 5, or with trauma caused by thermal injury, seizure, or diabetic ketoacidosis were excluded. Data obtained from the trauma registry included age, sex, BD, pH, trauma score,  Injury Severity Score (ISS),  probability of survival,  length of stay, length of intensive care unit stay (ICU days), complications, and survival. The criteria used to identify the complications of ARDS, renal failure, and coagulopathy were those adopted by the American Association for the Surgery of Trauma, as defined by Hoyt et al.  Using the same criteria, MOF was considered to be any combination of organ system dysfunction or failure, including two or more of the following: (1) pulmonary (ARDS or pulmonary failure), (2) cardiovascular (cardiogenic shock, circulatory failure), (3) gastrointestinal (hepatic failure, gastrointestinal hemorrhage), (4) coagulopathy or disseminated intravascular coagulation, (5) septicemia or sepsis-like syndrome, and/or (6) renal failure.
The hospital blood bank transfusion data base was used to determine packed red blood cell (PRBC) transfusion in the first 24 hours, total hospital PRBC transfusion, and total fresh frozen plasma transfusion. Medical record review was done to confirm and supplement registry data (e.g., missing data points, autotransfusion volumes).
Data are presented as mean +/- standard error of the mean. Statistical analysis was performed using analysis of variance (ANOVA), paired t test, chi2 test, linear regression analysis, and multivariate discriminate analysis (Jackknife technique) as noted. Statistical significance was attributed to p < 0.05.
From July 1, 1990 through August 31, 1995, 10,019 patients were admitted to the trauma service, with 5,264 patients meeting trauma registry criteria. Of these, 2,954 patients had an arterial blood gas obtained within 1 hour of admission and met the inclusion criteria. The mean patient age was 32.1 +/- 0.3 years (range, 5-88 years; median, 28 years), and 77% were male. The mechanism of injury was blunt in 71%.
Patients were stratified by BD category; "normal" (2 to -2, n = 1,377), "mild" (-3 to -5, n = 762), "moderate" (-6 to -9, n = 413) and "severe" (< or = -10, n = 279). An additional group of patients with "base excess" (>2, n = 123) was also analyzed.
Analysis of commonly used trauma scoring systems demonstrated changes with each BD category. The trauma score and revised trauma score showed significant decreases from the mild-to-moderate and moderate-to-severe categories (p < 0.001 and p < 0.001, respectively). The ISS increased in a stepwise fashion with increasingly severe BD (p < 0.001). Arterial pH decreased with each BD category (p < 0.001, all categories, Table 1), and linear regression analysis showed a strong correlation between pH and BD (r = 0.779, p < 0.001).
Changes in traditional clinical parameters were less pronounced. The heart rate increased across the BD categories from 90 in the "base excess" group to 100 in the "mild" group and 107 in the "severe" BD category. Although the changes in heart rate from the "normal" category were statistically significant (p < 0.001), the incremental changes in heart rate from the "mild" to "severe" category were not clinically discernible. Although mean systolic blood pressure decreased in a statistically significant fashion (p = 0.013 to p < 0.001) with each BD category, the mean systolic pressure was less than 90 mm Hg only in the severe category (Table 1).
BD category was indicative of resource utilization as reflected by transfusion requirements. The amount of PRBC transfused in both the first 24 hours after admission and overall during hospitalization increased significantly (p < 0.001) with each successive BD category (Table 2). Additionally, linear regression analysis revealed a significant relationship between BD and PRBC transfusion in the first 24 hours (r = -0.508, p < 0.001). The amount of fresh frozen plasma transfused also increased by BD category (Table 2), but the linear correlation was not as strong (r = -0.190, p < 0.001).
More significant, from a blood bank utilization perspective, was the proportion of patients that required any PRBC transfusion versus those who did not. The percentage of patients requiring transfusion increased with each successive BD category (p < 0.001, chi2 test) (Figure 1). PRBC transfusion was required in the first 24 hours in 72% of patients with BD < or = -6 and in only 18% of patients with BD > or = -5 (p < 0.001, chi sup 2 test).
The length of stay in both the ICU and the hospital increased as BD worsened (Table 3). The increase in ICU stay between the "moderate" and "severe" groups is demonstrated in the survivor subgroup (n = 2,531). The statistically significant differences in hospital length of stay also become more pronounced in the survivor subgroup, reflecting the increased early mortality of the "moderate" and "severe" BD groups.
Worsening BD was also indicative of an increased risk of developing complications. The incidence of ARDS, renal failure, coagulopathy, and MOF increased with worsening BD (Table 4). Multivariate analysis of the complications of coagulopathy, ARDS, and MOF was performed. In a best fit model, ISS was the most significant factor in predicting the development of complications. However, BD group added significantly to the predictive value of the model. This discriminate analysis using ISS and BD category demonstrated accuracy of 81.2% for coagulopathy, 77.1% for ARDS, and 76.7% for MOF. The probability curves for coagulopathy, ARDS, and MOF are shown in Figure 2, Figure 3 and Figure 4, respectively.
Of the 2,954 patients, 423 (14.3%) died. Mortality increased significantly (p < 0.001) with worsening BD category Figure 5). Decreased survival in the "base excess" group versus the "normal" group was attributed to the increased percentage (35% vs. 15%, p < 0.001, chi2 test) of patients with severe head injuries in the "base excess" group versus the "normal" BD group. Of note, the decrease in probability of survival with each successive BD group (p < 0.001) paralleled that of the observed survival (Table 5). There was no significant difference between predicted survival and observed survival in any BD category (p > 0.30).
Base deficit has been established as a reliable endpoint of resuscitation. In animal studies, BD was demonstrated to have a close correlation to serum lactate in shock and resuscitation.  BD has also been shown to have a strong relationship to indices of tissue oxygen utilization and to be reflective of tissue oxygen consumption, even in compensated shock. 
Compensated shock is usually a clinical diagnosis based on soft findings such as pallor, tachycardia, changes in sensorium, and decreased urine output. Extrinsic factors including ambient temperature, patient anxiety and agitation, and preexisting conditions may make clinical evaluation less reliable. This study confirms and expands upon previous findings that admission BD identifies compensated shock earlier than do changes in heart rate and blood pressure. Other clinical studies have also demonstrated the utility of BD in identifying compensated shock. [2,6]
The predictive value of BD has also been investigated. In a multifactorial analysis, BD was the best predictor of 29 parameters (including lactate, pH, and hemodynamic indices) of changes in blood volume in a canine model of hemorrhagic shock.  Similar results were obtained in comparing BD to established clinical predictors of intra-abdominal injury (hypotension, pelvic fracture, chest injury). [3,4] BD was shown to be the single most important predictor of intra-abdominal injury/hemorrhage in case-controlled multifactorial analysis. 
This investigation expands on previous work, but with some significant differences. In the initial clinical study, three BD categories were described: "mild" (2 to -5), "moderate" (-6 to -14), and "severe" (< or =-15).  A subsequent clinical study modified this by including a "normal" BD category (2 to -2), and altering the "mild" classification (-3 to -5).  This investigation has adjusted the classification for the "moderate" and "severe" categories. Clinical experience and mortality data from this report indicated that the "moderate" category was too broad. Previous investigations demonstrated the differentiation between "mild" and "moderate" occurred at a BD of -6. [2,4] In this study, significant changes in both injury severity and mortality occur at a BD of -6 and again at a BD of -10. The recognition of the more appropriate "cut-point" at a BD of -10 has led to the modification of the "moderate" (-6 to -9) and the "severe" (< or =-10) categories.
In this study, BD category was reflective of the calculated probability of survival. Furthermore, observed survival and predicted survival were essentially identical in each BD category. This differs from a previous investigation  where TRISS probability of survival was found to be a better discriminator of mortality than BD. However, in that investigation, the worst BD obtained in the first 24 hours was used as the admission value, leading to some skewing of the relationship of BD, probability of survival, and mortality.
Transfusion volumes increased both in the first 24 hours and overall with successive BD category. More importantly, the percentage of patients requiring any transfusion increased with each worsening BD category. Only 10% of patients in the "base excess" and 11% of patients in the "normal" groups received blood and 27% of patients in the "mild" group required blood. In contrast, 72% of the patients in the "moderate" (-6 to -9) and "severe" (< or =-10) groups received almost 4 and 8 units of blood, respectively, in the first 24 hours of hospitalization. Improved utilization of blood bank resources (including technician time) could be achieved by using admission BD. Patients with BD < or = -6 should undergo blood typing and cross-match (or use of type-specific blood). Patients with admission BD > -6 could be limited to a blood specimen sent for type and hold only, unless clinical circumstances dictate otherwise. Patients with "normal" or "base excess" values (BD > or = -2) could have a specimen drawn and held (but not typed or otherwise analyzed) because of the lower likelihood of blood utilization. This application of admission BD for directing blood testing and transfusion practices greatly expands upon the results of previous studies noting an association between BD and subsequent transfusion needs. [2,13]
BD category also effectively stratified patients for other types of resource utilization. Both ICU and hospital length of stay increased significantly with worsening BD category (Table 4). The major increase in both ICU and hospital length of stay occurs between the "mild" and "moderate" BD groups. However, when nonsurvivors were excluded, there is a significant increase in length of ICU stay between the "moderate" and "severe" groups as well.
Additionally, the incidence of shock-related complications (ARDS, MOF, renal failure, and coagulopathy) all increased with each successive BD category. In a study of postinjury MOF, BD and lactate were found to add substantially to the predictive value of the independent factors of age (>55 years), ISS (> 25) and transfusion (>6 units PRBC within 12 hours).  The findings of the discriminate analysis in this study show the additive value of BD category with ISS. Admission BD in conjunction with ISS therefore may help to identify patients at increased risk for coagulopathy, ARDS, and MOF. Heightened sensitivity to this increased risk may lead to more aggressive resuscitation and closer surveillance with invasive monitoring in these patients. Whether these measures can impact the severity of these complications and their outcomes in this patient group remains speculative.
Admission BD is reflective of both physiologic and anatomic injury scoring systems. This investigation has refined the BD categories from previous definitions; "normal" (2 to -2), "mild" (-3 to -5), "moderate" (-6 to -9), and "severe" (< or =-10). The requirement for PRBC transfusion and fresh-frozen plasma both increased with worsening BD category. Patients with a BD < or = -6 should undergo type and cross-match upon admission. Shock-related complications (ARDS, MOF, coagulopathy, and renal failure) increased as the BD became more severe. Early identification of patients at increased risk for these complications can be used to change strategies for resource utilization.
We thank the staff of the medical records department at Valley Medical Center, Fresno for their extraordinary assistance in this project.
1. Davis JW, Shackford SR, Hollbrook TL: Base deficit as a sensitive indicator of compensated shock and tissue oxygen utilization. Surg Gynecol Obstet 173:473, 1991
2. Davis JW, Shackford SR, Mackersie RC, et al: Base deficit as a guide to volume resuscitation. J Trauma 28:1464, 1988
3. Mackersie RC, Tiwary AD, Shackford SR, et al: Intraabdominal injury following blunt trauma. Arch Surg 124:809, 1989
4. Davis JW, Mackersie RC, Hollbrook TL, et al: Base deficit as an indicator of significant abdominal injury. Ann Emer Med 20:842, 1991
5. Siegel JH, Rivkind AI, Dala S, et al: Early physiologic predictors of injury severity and death in blunt multiple trauma. Arch Surg 125:498, 1990
6. Rutherford EJ, Morris JA Jr, Reed GW, et al: Base deficit stratifies mortality and determines therapy. J Trauma 33:417, 1992
7. Champion HR, Sacco WJ, Carnazzo AJ, et al: Trauma score. Crit Care Med 9:672, 1981
8. Baker SP, O'Neill B, Haddon W Jr, et al: The injury severity score: A method for describing patients with multiple injuries and evaluating emergency care. J Trauma 14:187, 1974
9. Boyd CR, Tolson MA, Copes WS: Evaluating trauma care: The TRISS method. J Trauma 27:370, 1987
10. Hoyt DB, Hollingsworth-Fridlund, Fortlage D, et al: An evaluation of provider-related and disease-related morbidity in a Level I university trauma service: Directions for quality improvement. J Trauma 33:586, 1992
11. Davis JW: The relationship of base deficit to lactate in porcine hemorrhagic shock and resuscitation. J Trauma 36:168, 1994
12. Waisman Y, Eichacker PQ, Banks SM, et al: Acute hemorrhage and validation of models to quantify blood loss. J Appl Physiol 74:510, 1993
13. Falcone RE, Santanello SA, Schulz MA, et al: Correlation of metabolic acidosis with outcome following injury and its value as a scoring tool. World J Surg 17:575, 1993
14. Sauaia A, Moore FA, Moore EE, et al: Early predictors of postinjury multiple organ failure. Arch Surg 129:39, 1994