OTA 2005 Posters

Scientific Poster #12 Polytrauma/Injury Prevention

:The Predictive Value of Base Deficit for Adult and Pediatric In-Hospital Mortality

Jason C. Fanuele, MD; E.W. Weber, MD; R.V. Cantu, MD;

K.F. Spratt, PhD; Kenneth J. Koval, MD (n-all authors);

Dartmouth Hitchcock Medical Center, Lebanon, New Hampshire, USA

Purpose: Base deficit (BD) is an accurate measure of end organ perfusion readily available from an arterial blood gas. The purpose of this study was to: (1) confirm previous work demonstrating the prognostic value of initial BD for in-hospital mortality in pediatric and adult trauma populations; (2) demonstrate the utility of age stratification in the pediatric population to guide treatment in this medically heterogeneous population; and (3) compare the ability of BD and Injury Severity Score (ISS) to predict in-hospital mortality across age groups.

Methods: This review included all patients from the National Trauma Data Bank (NTDB-Version 4.0) between the ages of 0 and 100 years (pediatric defined as 0 to 17 years) with valid BD information, ISS, and discharge status. Logistic regression techniques were used to: (1) verify the relationship between BD and mortality in the adult population reported from previous NTDB data; (2) determine appropriate BD and ISS "cut point" scores for predicting patient mortality; and (3) evaluate the predictive ability of the age-specific clinical interpretation of BD and ISS for patient in-hospital mortality.

Results: A total of 76,181 adult and 10,823 pediatric patients met study inclusion criteria. In adults, logistic regression modeling indicated that mortality was 5.86 times (99% CI, 5.42 - 6.87) more likely for patients with higher risk (-5.0 to -30.0) compared with patients with lower risk (+3.0 to -4.9) BD levels. By age groups, mortality was reliably more than doubled (P <0.01), with (1) BD -7 for patients 0 to 1 years old, (2) BD < -3 for patients 2 to 5 years old, (3) BD -9 for patients 6 to 12 years old, (4) BD -7 for patients 13 to 15 years old, and (5) BD -5 for patients 16 to 17 years old. Applying these cut scores with similarly derived ISS cut scores across the entire pediatric sample, BD was a more accurate predictor of mortality that ISS, with diagnostic confidence demonstrating an improvement ratio (successful model mortality prediction rate/sample mortality rate) of 2.62 for BD compared to 1.97 for ISS.

Conclusions: (1) BD was a better predictor of in-hospital mortality than ISS across all age categories. (2) Age stratification had different values of BD that affected mortality prediction. (3) These results confirmed previous findings that initial BD was predictive of in-hospital mortality in adults and pediatric patients, whichgiven the less than 10% incidence of getting BD in this trauma databasesuggests that this diagnostic test is underutilized in trauma settings.