Journal of Pediatric Surgery 49 (2014) 1678–1682

Contents lists available at ScienceDirect

Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg

Initial hematocrit predicts the use of blood transfusion in the pediatric trauma patient☆,☆☆,★ Casey J. Allen a, Jun Tashiro b, Evan J. Valle a, Chad M. Thorson b, Sherry Shariatmadar c, Carl I. Schulman a, Holly L. Neville b, Kenneth G. Proctor a, Juan E. Sola b,⁎ a b c

Division of Trauma Surgery, DeWitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine, Miami, FL, USA Division of Pediatric Surgery, DeWitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine, Miami, FL, USA Division of Hematology, DeWitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine, Miami, FL, USA

a r t i c l e

i n f o

Article history: Received 14 October 2014 Accepted 15 October 2014 Key words: Hematocrit Transfusion Children Adolescents Kids

a b s t r a c t Purpose: Initial hematocrit (Hct) is generally not considered a marker of acute blood loss because it is assumed that physiologic response of fluid conservation to hemorrhage is delayed. We challenged this idea by theorizing that admission Hct correlates with conventional signs of shock and predicts the use of blood transfusion during resuscitation of pediatric trauma patients. Methods: Data from 1928 pediatric admissions (b 18 years) at a Level I trauma center (2000–2012) were compared using standard statistical analyses and logistic regression modeling to identify factors associated with blood transfusion during initial trauma resuscitation. Results: Overall mortality rate was 3.5%, with a transfusion rate of 10.7%. Factors significantly associated with transfusion were initial Hct, Glasgow Coma Score, base deficit, and injury severity score (all p b 0.001). Initial Hct is a stronger predictor for transfusion (area under receiver operator curve (AUC: 0.728) compared to agespecific tachycardia (AUC: 0.689), age-specific hypotension (AUC: 0.673), and altered mental status (AUC: 0.654)). On multivariate analysis, initial Hct was an independent predictor (OR [95% CI]: 2.94 [1.56, 5.52]) along with hypotension (6.37 [2.95, 13.8]), base deficit (4.14 [1.38, 12.4]), and tachycardia (3.07 [1.62, 5.81]). Conclusions: Initial Hct correlates significantly with conventional signs of shock and is a strong independent predictor for blood transfusion with better predictability than other clinical factors. © 2014 Elsevier Inc. All rights reserved.

The use of blood transfusions during initial resuscitation of the pediatric trauma patient is based on suspicion for acute blood loss and potential for ongoing hemorrhage. Blood pressure, heart rate, and mental status are used to evaluate hemodynamic status. However, normal respiratory rate, heart rate (HR), and systolic blood pressure (SBP) differ from adults and vary within age groups in the pediatric population [1]. Thus, with a wide range of normal physiologic values, the use of vital signs may be difficult during initial assessment. Furthermore, pediatric patients, typically have an increased physiologic reserve and can maintain normal vital signs until shock is severe. In children, the standard physiologic response to hypovolemia is peripheral vasoconstriction and reflex tachycardia; this may mask significant hemorrhagic injury, because children can compensate for up to a 25% loss of circulating blood volume with minimal external signs [2].

☆ Portions of the data presented at the 2014 Templeton Pediatric Trauma Symposium. ☆☆ Supported in part by: Grants No. N140610670 from the Office of Naval Research and No. 09078015 from U.S. Army Medical Research & Material Command. ★ None of the authors have declared conflicts of interest. ⁎ Corresponding author at: Division of Pediatric Surgery, DeWitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine, 1120 NW 14th Street, Suite 450K, Miami, FL 33136, USA. Tel.: +1 305 243 5072. E-mail address: [email protected] (J.E. Sola). http://dx.doi.org/10.1016/j.jpedsurg.2014.10.044 0022-3468/© 2014 Elsevier Inc. All rights reserved.

Initial signs of shock can be subtle in a child; perhaps manifested only by lethargy. Also, during the critical period of evaluation, a frightened child may be emotionally labile (which may be confused with altered mental status) and/or tachycardic from fear and pain (which may be confused with hemodynamic instability). By the time hypotension has developed, a child may have already lost approximately 40% of their circulating blood volume. Traumatic arrest is imminent if treatment is delayed [2]. Thus, “normal” values for vital signs should not necessarily make one feel more secure about the child's volume status. Traditionally, initial hematocrit (Hct) has not been regarded as an accurate marker of blood loss and shock in trauma patients [2]. Based on the fact that “patients bleed whole blood” and on the assumption that compensatory mechanisms for fluid retention and absorption into the vascular space are relatively slow, surgeons are traditionally taught that initial Hct remains within the normal range soon after hemorrhage [3,4]. Some still believe that Hct can be near normal despite severely contracted blood volume and therefore is not useful during initial evaluation [5]. On the other hand, recent reports conclude that hemorrhageinduced hemodilution occurs within minutes and admission Hct is one of its strongest indicators [6]. Thus, there is some question whether fluid actually shifts rapidly after bleeding. Regardless, the use of admission Hct as a clinical tool to predict blood transfusion requirements has not been assessed in the pediatric trauma population.

C.J. Allen et al. / Journal of Pediatric Surgery 49 (2014) 1678–1682

The purpose of this study is to evaluate the association between initial Hct and need for blood transfusion during initial resuscitation of the pediatric trauma patient. We hypothesize that initial Hct correlates with clinical signs of shock as well as blood transfusion requirements. 1. Methods This retrospective analysis of trauma registry data over a 13 year period (Jan 2000–Dec 2012) was approved by the University of Miami Miller School of Medicine Institutional Review Board. Inclusion criteria were all trauma patients b18 years of age. Exclusion criteria were those pregnant, incarcerated, or those not admitted to either the trauma or pediatric surgery service. Clinical measures, such as vital signs, type of injury, initial Hct, and base deficit, were gathered to identify predictors for the use of blood transfusion during the initial resuscitation period. Continuous data are reported as mean ± standard deviation and are compared using a t-test or Mann–Whitney U test, as appropriate. Categorical variables were compared using a chi-square or Fisher's exact test, as appropriate. Calculations of negative predictive value (NPV), positive predictive value (PPV), sensitivity, and specificity were used to determine the predictive capability of categorical variables for the use of blood transfusion products. Pearson correlation coefficient (r) tests were performed to detect clinical characteristics correlated with increasing Hct and base deficit values. In addition, receiver-operator curves (ROC) were used to determined the predictive capability of individual variables; presented as area under the ROC (AUROC), where 1.0 is perfect predictability and 0.5 represents the null hypothesis. The results of the univariate analyses were used to identify variables for inclusion in a binary logistic regression model. Statistical significance was determined at alpha level 0.05. Statistical analyses were performed using SPSS version 21 (IBM Corporation; Armonk, NY). 2. Results The sample size comprised 1928 patients with 3.5% mortality and length of stay (LOS) of 7 ± 12 days. The cohort was 70% male, with a mean age of 11 ± 6 years old, 76% blunt injury, with an injury severity score (ISS) 13 ± 12. Patients were transferred from the resuscitation area to a regular ward (45%), intensive care unit (31%), or required emergent surgery (24%). See Table 1. The transfusion rate was 10.7% in the initial trauma resuscitation period. Patients requiring transfusion were of older age, had higher ISS, less likely sustained blunt force trauma, had a more severe base deficit, lower Hct, more frequent age-specific tachycardia, hypotension, lower Glasgow Coma Scale (GCS), higher rates of urgent operation, higher mortality, and longer LOS (all, p b 0.001). See Table 2.

Table 2 Comparison of patients requiring vs. not requiring the use of transfusion products. Category

No transfusion (n = 1722)

Transfusion (n = 206)

p Value

Age (y) Male Blunt ISS Base deficit Hematocrit (%) Tachycardia Hypotension GCS Urgent operation LOS (d) Death

11 ± 6 70% 77% 11 ± 10 2±4 37 ± 5 28% 3% 14 ± 3 18% 6±8 1.4%

14 ± 4 70% 57% 31 ± 15 7±6 33 ± 7 66% 35% 10 ± 5 72% 20 ± 25 21.9%

b0.001 NS b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001

Continuous measures are compared using t-test or Mann–Whitney U test as appropriate for equal or unequal variance. Categorical data are compared using chi-square or Fisher's exact test as appropriate.

Admission Hct significantly correlated with total amount of blood transfusion required (r = −0.193), ISS (r = −0.202), GCS (r = 0.218), age-specific hypotension (r = −0.173), age-specific tachycardia (r = −0.031), LOS (r = −0.174), and mortality (r = −0.241) (all p b 0.001). Initial base deficit also significantly correlated with total amount of blood transfusion required (r = 0.364), ISS (r = 0.377), GCS (r = −0.334), age-specific hypotension (r = 0.343), age-specific tachycardia (r = 0.241), LOS (r = 0.137), and mortality (r = 0.447) (all p b 0.001). See Table 3. Initial Hct was the strongest predictor for the use of transfusion (AUC: 0.728). Other associated factors were age-specific tachycardia (AUC: 0.689), hypotension (AUC: 0.673), and altered mental status (AMS) (AUC: 0.654). When assessing the predictive capability of clinical factors associated with the use of transfusion, an initial Hct b27% had a specificity of 98%, while Hct b24% had a specificity approaching 99%. Similarly, hypotension (97%) and base deficit N7 (95%) were extremely specific factors as well. For all factors, sensitivity for predicting transfusion was relatively low, with the exception of base deficit greater than 0 (94%). See Table 4. Multivariate analysis identified several clinical characteristics as significant predictors of transfusion. Hypotension was the greatest independent prognostic indicator for transfusion (OR: 6.37 [2.95, 13.8]), followed by base deficit (4.14 [1.38, 12.4]), tachycardia (OR: 3.07 [1.62, 5.81]), and Hct b36% (2.94 [1.56, 5.52]). AMS was not a significant indicator for transfusion on logistic regression (1.79 [0.95, 3.38]). See Fig. 1.

3. Discussion The major new findings from this study in pediatric trauma patients are that, in contrast to conventional teaching, Hct predicts the need for blood transfusion better than age specific hypotension, tachycardia, and AMS. After controlling for other variables, initial Hct remains a

Table 1 Demographic characteristics of cohort. Characteristics Age (y) Injury Severity Score LOS (d) Transfusion Mortality Gender Male Female Type Blunt Penetrating Disposition from ED Floor ICU OR

1679

11 ± 6 13 ± 12 7 ± 12 10.7% 3.5% 70% 30% 76% 24% 45% 31% 24%

Continuous values are shown as mean ± standard deviation. Total cases in cohort = 1928.

Table 3 Pearson correlation coefficient (r) tests were performed between clinical categories and increasing hematocrit or base deficit. Category

Hematocrit

p Value

Base deficit

p Value

PRBC (units) ISS GCS Hypotension Tachycardia LOS Mortality

−0.193 −0.202 0.218 −0.173 −0.031 −0.174 −0.241

b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001

0.364 0.377 −0.334 0.343 0.241 0.137 0.447

b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001

Respective p-values are shown.

1680

C.J. Allen et al. / Journal of Pediatric Surgery 49 (2014) 1678–1682

Table 4 Predictive values for clinical characteristics. Category

Sensitivity

Specificity

Tachycardia Hypotension AMS Hct b 36 Hct b 27 Hct b 24 BD N 0 BD N 5 BD N 7

66% 35% 60% 66% 14% 7% 94% 54% 29%

72% 97% 72% 69% 98% 99% 30% 88% 95%

Sensitivity and specificity values are used to examine whether the test proposed correctly predicts the use vs. nonuse of transfusion products.

significant independent predictor for the need for blood transfusion during initial trauma resuscitation. Altogether, this challenges conventional wisdom but supports and extends our previous findings. During the primary survey of trauma patients, early and rapid assessment of bleeding and the need for blood volume replacement is crucial. The limitations of conventional vital signs and laboratory values are well established in pediatric trauma patients. Our findings suggest that SBP has good specificity (97% for age specific hypotension), but lacks sensitivity (35%) for recognizing the need for blood transfusion. Tachycardia is also not always a reliable indicator, suffering a lower specificity (72%). In addition, several trauma guidelines use these conventional vital signs, but not Hct, as transfusion criteria. Conventional teaching holds that Hct remains stable for several hours after trauma because “patients bleed whole blood” and compensatory mechanisms for fluid shifts have not yet taken place [3,5,7]. However, there are multiple controlled studies in animals [8–11] and humans [12–15] showing an almost immediate drop in Hct following hemorrhage. As the initial blood draw at our trauma center is typically performed within 5 minutes upon arrival, sent to a laboratory with results usually available within 15 minutes, our study's findings support the conclusion that the admission Hct implies that fluid shifts from the interstitial space into the vasculature more rapidly than current resuscitation guidelines and surgical textbooks would indicate. Knottenbelt [16] published one of the first clinical studies directly investigating hemoglobin in trauma. Of 31 adult patients with a presenting hemoblogin ≤ 8 g/dL, 42% died of hypovolemia compared to only 2.6% of those with higher levels. Snyder [17] found that adult patients with an initial Hct ≤ 35% had a greater need for emergent operation (41% vs. 7%) and higher mortality (10% vs. 1%). Bruns et al. [18] showed a significant correlation between Hct, conventional vital signs (SBP, HR),

laboratory markers (pH, base deficit), and need for blood products in the first 24-hours in the adult population. Our previous studies have corroborated these findings, with a significant correlation between the initial Hct and EBL in adult patients requiring emergent trauma operations [6]. Because of the wide range of normal physiological values and the potential unreliability of conventional vital signs within the various age groups of the pediatric population, we examined the utility of the admission Hct in pediatric trauma patients. Our current results show that initial Hct is significantly correlated with acidosis, agespecific hypotension, age-specific tachycardia, multiple injury severity indicators (ISS, RTS), LOS, and mortality. This is consistent with our prior study in adults that demonstrated significant correlation in trauma patients between the initial Hct and tachycardia, systolic hypotension, low pH, and a higher base deficit [6,18]. The acidosis is likely the result of systemic hypoperfusion alone, since humans tolerate significant isovolemic hemodilution without changes in systemic lactate or increases in base deficit [19]. Our results also show that initial Hct value strongly correlates with the amount of PRBCs transfused during the initial resuscitation period. These findings agree with previous authors who demonstrated significant correlations between the initial Hct and transfusion of PRBCs as well as overall crystalloid requirements in adults [17]. A joint task force of the Eastern Association for the Surgery of Trauma and the Society of Critical Care Medicine has developed clinical practice guidelines for PRBC transfusion in adult trauma and critical care [20]. There is level 1 evidence to support emergent transfusions for patients with hemorrhagic shock, and in those with acute hemorrhage and hemodynamic instability and/or inadequate oxygen delivery. Level 2 evidence suggests that the use of Hct as a transfusion trigger should be avoided, and decisions should be based on intravascular volume status, cardiopulmonary physiology, evidence of shock, and duration and extent of anemia [20]. Several studies have demonstrated poor sensitivity of Hct for identifying bleeding or the need for an emergent operation [17,18,21,22]. Therefore, lack of depression in the initial values does not rule out the presence of significant blood loss or ongoing bleeding. These results do not indicate that the initial Hct is useful to determine whether the patient is actively hemorrhaging; it only reflects that bleeding has already occurred. In the absence of other clinical signs, the practical significance of a single Hct value is questionable. In addition, the confounding influence of intravenous fluid use makes interpretation of a single Hct value difficult [13,14,17], especially because prehospital fluids are often unknown. Patients presenting to our trauma center have a mean transport time of b 17 min, and it is unlikely that low initial Hct can be explained entirely by preadmission fluids, but we cannot absolutely rule out this possibility.

OR [CI]

AMS

1.79 [0.95, 3.38]

Tachycardia

3.07 [1.62, 5.81]

Hypotension

6.37 [2.95, 13.76]

Abnormal BE (

Initial hematocrit predicts the use of blood transfusion in the pediatric trauma patient.

Initial hematocrit (Hct) is generally not considered a marker of acute blood loss because it is assumed that physiologic response of fluid conservatio...
228KB Sizes 1 Downloads 9 Views