Early leukocytosis in trauma patients: what difference does it make?

SBAS MEETING

Early Leukocytosis in Trauma Patients: What Difference Does It Make? David C. Chang, MPH, Edward E. Cornwell III, MD, Judith Phillips, RN, Jane Paradise, and Kurtis Campbell, MD Division of Adult Trauma, Department of Surgery, Johns Hopkins School of Medicine, Baltimore, Maryland OBJECTIVE: To determine the association of the admission white blood cell count in trauma patients with demographics, severity and mechanism of injury, and need for therapeutic intervention. METHODS: Evaluation of prospectively collected registry

data (admissions to a Level I trauma center in 2001). Differences in mean white blood cell count on admission were evaluated with t tests. Multiple linear regressions were performed with forward stepwise selection of variables. RESULTS: Of the 882 patients admitted for greater than 24

hours, white blood cell count was available for 786. Variations in white blood cell count were noted on bivariate analysis among different races, injury mechanisms and severities, Glascow Coma Scores, blood pressures, and between patients requiring early transfusions versus those who did not. No difference was noted between patients who went to the operating room in the first 24 hours versus those who did not, or for patients who died in the hospital. On multiple linear regression analyses, only ISS greater than 15, GCS less than or equal to 8, and white race were associated with increases in white blood cell count. Leukocytosis was found not to be associated with mechanism of injury, specific organ injury, shock on admission, or the need for transfusion or surgery. CONCLUSION: Variations in white blood cell count in

trauma patients are associated with race and injury severity, but they are not beneficial in predicting the need for volume resuscitation, transfusion, or surgery. (Curr Surg 60:632-635. © 2003 by the Association of Program Directors in Surgery.) KEY WORDS: trauma, leukocytosis, white blood count, in-

jury

INTRODUCTION Costs and resource utilization of trauma care have come under scrutiny across the country. Trauma system and trauma centers Correspondence: Inquiries to Edward E. Cornwell III, MD, The Johns Hopkins Hospital, 600 North Wolfe Street, Osler 625, Baltimore, Maryland 21287; fax: (410) 955-1884; e-mail: [email protected]

632

are threatened with collapse due to the financial burdens represented by the unreimbursed cost of care.1 Laboratory profile on trauma patients typically includes a complete blood count (CBC). Some have suggested white blood cell count contributes to prognostication of outcome in trauma.2 However, the real-time clinical contribution of this information is uncertain. It is well known that leukocytosis is associated with physiologic stress, due to catecholamine-induced demargination of white cells.3-5 The objective of this study is to determine if increased white blood cell count is correlated with demographics, mechanism and severity of injury, the need for therapeutic intervention, or outcome in trauma patients.

MATERIALS AND METHODS This is a retrospective evaluation of prospectively collected information in the registry of a Level I trauma center regarding “major” trauma admissions for patients age greater than or equal to 15 years in 2001 (Table 1). Patient information included gender, race, mechanism of injury (blunt versus penetrating trauma), Injury Severity Score (ISS), head Abbreviated Injury Scale (AIS) score, systolic blood pressure (SBP), Glasgow Coma Score (GCS), units of blood transfused in the first 24 hours, need for surgery, and mortality. The white blood cell count was measured from blood samples taken immediately upon admission. As the majority of the trauma patient population served by this institution comes from within a 5-mile radius of the hospital with short transport times, and with very few interhospital transfers (3.3%), these samples were typically drawn shortly after injury.6 The data analysis was performed in Intercooled Stata version 6.0 (Stata Corporation, College Station, Texas). Comparisons of means were performed using Student’s t test; comparisons of medians were performed using Kruskal-Wallis test; comparisons of proportions were performed using chi-square test; and multiple linear regressions were performed with forward stepwise selection of variables, with significance set at p ⫽ 0.05.

CURRENT SURGERY • © 2003 by the Association of Program Directors in Surgery Published by Elsevier Inc.

0149-7944/03/$30.00 doi:10.1016/j.cursur.2003.07.011

TABLE 1. “Major” Trauma Admission Criteria I. Hemodynamic Instability SBP ⬍ 90 P ⬎ 120 or ⬍50 RR ⬎ 30 or ⬍10 II. Mechanism of Injury - Penetrating Injury to: Head, neck, torso, isolated extremity (if bleeding or obvious fracture) - All Shotgun or Buckshot Injuries - Fall ⬎ 15 feet - Assault with major long bone fracture - Pedestrian struck - Motor vehicle crash—if co-occupant death, ejection, prolonged extrication - All motorcycle and moped crashes - Head/spine trauma—GCS ⱕ 13 or neurologic deficit - Burns—⬎15% TBSA or with airway compromise - Electrical—high voltage or with sustained arrhythmias III. Specific Injuries - Major vascular injury - Open fractures—excluding hands and feet - Traumatic pneumothorax - Traumatic amputation or major degloving injury - Depressed skull fracture - Fracture of: (any one) Cervical, thoracic, or lumbar spine First or second ribs Clavicle (if secondary to fall from height or motor vehicle crash) Sternum Thoracic and/or lumbar spine Pelvis or sacrum (even if stable) Femur (excluding low velocity hip fractures in elderly patients) - Hip Dislocation—associated with blunt trauma - Knee dislocation - Scapular—thoracic dissociation SBP ⫽ systolic blood pressure; P ⫽ pulse; RR ⫽ respiratory rate; GCS ⫽ Glasgow Coma Score; TBSA ⫽ total body surface area.

RESULTS Table 2 presents the bivariate analysis of patients’ mean white blood cell count as they relate to demographics, injury characteristics, hospital course, and mortality. Significant differences in mean white blood cell count were seen between white and black patients, between patients with blunt versus penetrating trauma, between ISS greater than 15 and ISS less than 15, between GCS greater than 8 and GCS less than or equal to 8, between normotensive and hypotensive patients, between patients with and without splenic injuries, and between patients requiring transfusion in the first 24 hours versus those who did not. No difference in mean white blood cell count was associated with presence or absence of bowel or liver injury, or for the need for surgical intervention in the first 24 hours, or with survival. In addition, white blood cell count at the extremes was associated with patient lengths of stay (median LOS ⫽ 2 days for white blood cell count greater than 12,000 or less than 4000, versus median LOS ⫽ 1 day for others, p ⬍ 0.01), but not with mortality.

On multiple linear regression with forward stepwise selection, ISS greater than 15, GCS less than or equal to 8, and white race were the only factors that emerged with an association to higher white blood cell count. Several clinical parameters associated with differences in white blood cell count on bivariate analysis did not emerge as independent factors on multivariate analysis, including mechanism of injury, splenic injury, hypotension on admission, or need for transfusion.

DISCUSSION There have been variable conclusions in the literature regarding the value of measuring the white blood cell count for trauma patients. Questions have been raised because of the well-described concept of catecholamine release and demargination of leukocytes upon major injury and surgical stress.3-5 Variations in interpretations of the literature may well be explained by drawing a distinction between the statistical significance and the clinical significance of differing admission white blood cell count. Malone et al analyzed data on over 9500 admissions at a Level I trauma center in order to validate previous findings that the Systemic Inflammatory Response Syndrome (SIRS) score predicted mortality and length of stay in trauma patients.2,7 The score is calculated by assigning a point for abnormal values in each of four components: white blood cell count (⬎12,000 or ⬍4000), temperature (⬎38 or ⬍36), heart rate (⬎90), and respiratory rate (⬎20). A score of greater than one was predictive for increased mortality and length of stay. Analysis of the specific components of the SIRS score revealed that hypothermia was the best predictor of mortality, whereas leukocytosis correlated with increased length of stay (as it did in this study). Although these are interesting prognosticators, a clinician is more likely to consider the updated white blood cell count value as the day of discharge approaches, rather than the white blood cell count on admission. The association of white blood cell count and outcome of traumatic brain injury is well described. The use of GCS less than or equal to 8 cutoff was selected in this analysis to coincide with the severity of head injury where the guidelines for diagnostic and therapeutic interventions are best delineated.8 Rovlias et al reported that admission white blood cell count correlated with GCS, pupillary reaction, presence of subarachnoid hemorrhage, and unfavorable outcomes in patients with traumatic brain injury.9 Souter et al used serial white blood cell count in order to derive their conclusion that it was the magnitude of change from the admission white blood cell count that predicted the development of delayed intracranial hypertension among head injured patients.10 In this regard, the admission white blood cell count is contributory mainly to the extent that it varies from the subsequent white blood cell count. Our finding of a correlation between admitting white blood cell count and GCS on multivariate analysis is consistent with these two reports. In a study of 46 patients with isolated intestinal injury after

CURRENT SURGERY • Volume 60/Number 6 • November/December 2003

633

TABLE 2. Basic Demographic Information, Injury Characteristics, and Hospital Course of Study Patients and Their White Blood Cell Count Data number (%)

mean WBC (SD)

p value

All

786

10.0 (4.7)

—

Men Women

602 (76.6%) 184 (23.4%)

9.9 (4.5) 10.4 (5.4)

0.16

Blacks Whites

569 (72.3%) 180 (23.0%)

9.5 (4.1) 11.5 (6.2)

<0.01

Blunt Penetrating

455 (57.6%) 331 (42.1%)

10.5 (5.4) 9.3 (3.5)

<0.01

ISS⬎15 ISS⬍15

155 (19.7%) 621 (79.0%)

11.9 (6.6) 9.5 (4.0)

<0.01

GCSⱕ8 GCS⬎8

56 (7.1%) 693 (88.2%)

13.0 (7.4) 9.8 (4.4)

<0.01

SBP⬍90 SBPⱖ90

30 (3.8%) 741 (94.3%)

11.8 (8.4) 9.9 (4.5)

0.04

Hollow viscus injury (HVI) No HVI

27 (3.4%) 759 (96.6%)

10.0 (5.0) 10.0 (4.7)

0.97

Splenic injury No splenic injury

24 (3.1%) 762 (96.9%)

12.0 (4.6) 9.9 (4.7)

0.04

Liver injury No liver injury

33 (4.2%) 753 (95.8%)

10.3 (4.7) 10.0 (4.7)

0.70

Transfusion first 24 hours No transfusion

78 (9.9%) 708 (90.1%)

11.7 (6.9) 9.8 (4.4)

<0.01

OR first 24 hours Non-operative

176 (17.4%) 610 (77.6%)

10.3 (5.4) 9.9 (4.5)

0.36

Deaths Lived

22 (2.8%) 764 (97.2%)

10.3 (4.0) 10.0 (4.7)

0.75

blunt trauma, Harris et al suggested leukocytosis was an important indicator of injury in the subset of patients (“occult” injury) whose need for surgery was not immediately apparent.11 Their findings, however, are not inconsistent with our observations, because only 41% of their patients who ultimately had leukocytosis had increased white blood cell count on their initial measurements. Furthermore, even when serial white blood cell counts are considered, leukocytosis carried a 15% falsenegative rate, and only a 45% positive predictive value. Fakhry et al analyzed the database of over 250,000 trauma admissions and found that mean initial white blood cell count was not significantly different for patients with perforated small bowel injury (SBI, 14,900), nonperforated SBI (15,100), and no SBI (14,600).12 We similarly found that the admission white blood cell count did not aid in the decision for emergency abdominal surgery in any patient during our study period. The values were frequently not available at the time of the trip to the operating room. Although there was no race-based difference in outcome, our finding of racial differences in admission white blood cell count is intriguing and is consistent with reports in the literature. It has been found that the white blood cell count in healthy subjects is significantly lower in blacks than in whites,13 and this is so even after adjusting for the effects of gender, age, height, 634

body mass index, alcohol use, and smoking.14 Our presumption is that black patients began with a lower pre-injury white blood cell count. The final arbiter of the value of the admission white blood cell count should be the contribution that the white blood cell count makes to early clinical judgment in real time. Consider the 24-year-old African-American man in the operating room undergoing emergent laparotomy and receiving blood within 15 minutes of arrival for an abdominal gunshot wound. Although our data base allows detection of a statistically significant difference among small variations of the white blood cell count, a clinician is hard pressed to assign great meaning to differences of white blood cell count that may vary from the 9000 s to the 11,000 s (Table 2). In the clinical example cited, the admitting white blood cell count would not be known at the time of surgery, and indeed, it is not necessary to determine the patient’s race, mechanism of injury, ISS (which is ultimately calculated on discharge), GCS, blood pressure, or need for transfusion—parameters that on bivariate analysis were found to correlate with variations in white blood cell count. Given our finding on multiple linear regression analysis of an association between white blood cell count and GCS less than or equal to 8, combined with the findings of Souter et al of a correlation between changes in white blood cell count over time

CURRENT SURGERY • Volume 60/Number 6 • November/December 2003

and the development of delayed intracranial hypertension, we maintain an interest in measuring admission white blood cell count only among comatose patients. However, even among these patients, clinical decisions are driven by serial neurologic examinations and computed tomography scan findings. It is difficult to justify measuring admission white blood cell count in all other cases.

memory status following isoproterenol infusion. J Neuroimmunol. 2000;102:137-144. 6. Chang DC, Cornwell EE, Phillips J, Baker D, Yonas M,

Campbell K. Community characteristics and demographic information as determinants for a hospital-based injury prevention outreach. Arch Surg. In press. 7. Napolitano LM, Ferrer T, McCarter RJ Jr, Scalea TM.

Systemic inflammatory response syndrome score at admission independently predicts mortality and length of stay in trauma patients. J Trauma. 2000;49:647-652.

CONCLUSION On multivariate analysis, variations in white blood cell count are associated with race, ISS, and GCS less than or equal to 8. It is not beneficial in predicting the need for volume resuscitation, transfusion, or surgery. Restricting admission white blood cell count evaluation to comatose patients (GCS ⱕ 8) with head injury would be a more judicious use of resources consistent with published studies.

8. Brain Trauma Foundation. Indications for intracranial

pressure monitoring. J Neurotrauma. 1996;13:667-679. 9. Rovlias A, Kotsou S. The blood leukocyte count and its

prognostic significance in severe head injury. Surg Neurol. 2001;55:190-196. 10. Souter MJ, Andrews PJ, Pereirinha MR, Signorini DF,

Jones PA, Miller JD. Delayed intracranial hypertension: relationship to leukocyte count. Crit Care Med. 1999;27(1):177-181.

REFERENCES 1. Salganik M. State’s trauma system seeks infusion of $15

million a year. Baltimore Sun. January 15, 2003.

11. Harris HW, Morabito DJ, Mackersie RC, Halvorsen RA,

2. Malone DL, Kuhls D, Napolitano LM, McCarter R, Sca-

lea T. Back to basics: validation of the admission systemic inflammatory response syndrome score in predicting outcome in trauma. J Trauma. 2001;51:458-463.

Schecter WP. Leukocytosis and free fluid are important indicators of isolated intestinal injury after blunt trauma. J Trauma. 1999;46:656-659. 12. Fakhry SM, Watts DD, Luchette FA. Current diagnostic

3. Kuhlwein EC, Irwin MR, Ziegler MG, Woods VL,

Kennedy B, Mills PJ. Propranolol affects stress-induced leukocytosis and cellular adhesion molecule expression. Eur J Appl Physiol. 2001;86:135-141.

approaches lack sensitivity in the diagnosis of perforated blunt small bowel injury: analysis from 275,557 trauma admissions from the EAST multi-institutional HVI trial. J Trauma. 2003;54:295-306.

4. Landmann RM, Muller FB, Perini C, Wesp M, Erne P,

13. Reed WW, Diehl LF. Leukopenia, neutropenia, and re-

Buhler FR. Changes of immunoregulatory cells induced by psychological and physical stress: relationship to plasma catecholamines. Clin Exp Immunol. 1984;58(1):127-135.

duced hemoglobin levels in healthy American blacks. Arch Intern Med. 1991;151:501-505. 14. Schwartz J, Weiss ST. Host and environmental factors

5. Mills PJ, Goebel M, Rehman J, Irwin MR, Maisel AS.

Leukocyte adhesion molecule expression and T cell naive/

CURRENT SURGERY • Volume 60/Number 6 • November/December 2003

influencing the peripheral blood leukocyte count. Am J Epidemiol. 1991;134:1402-1409.

635