ORIGINAL ARTICLE
Injury patterns associated with hypotension in pediatric trauma patients: A national trauma database review Alison R. Gardner, MD, MS, Debra I. Diz, PhD, Janet A. Tooze, PhD, MPH, Chadwick D. Miller, MD, MS, and John Petty, MD, Winston-Salem, North Carolina
Hypotension after trauma is most commonly assumed to be hemorrhagic, or hypovolemic, in origin. However, hypotension may occur in pediatric patients with isolated head injury, challenging accepted tenets of trauma care. We sought to quantify the contribution of head injury to the development of hypotension after pediatric trauma. METHODS: This is a retrospective cohort analysis using the National Trauma Data Bank registry 2009. Children aged 0 to 15 years were classified by injury pattern sustained during trauma using discharge diagnosis International Classification of Diseases, Ninth Revision, codes into isolated head, hemorrhagic, spinal cord, or other injury type. The primary outcome was hypotension for age at arrival to the emergency department. Risk of hypotension was estimated and compared by injury pattern using absolute and relative risks (RRs) stratified by age group (0Y4 years, 5Y11 years, 12Y15 years). RESULTS: Rates of hypotension ranged from 1.8% to 2.3% by age, with the highest incidence in the 12- to 15-year group. The RR of hypotension from isolated head injury (RR, 2.5; 95% confidence interval, 2.0Y3.2 vs. other) was not significantly different from the RR for hemorrhagic injury (RR, 2.7; 95% confidence interval, 2.1Y3.5 vs. other) in the 0- to 4-year-old group. For the older age groups, the RR of hypotension from isolated head injury was significantly lower than from hemorrhagic injury. CONCLUSION: Hypotension occurs after isolated head injury in children, and the risk of hypotension is as great as hemorrhagic injuries in children aged 0 to 4 years. This finding should now lead us to confirm whether a cause-effect relationship exists and, if so, isolate the responsible mechanism. In turn, this could reveal an opportunity to tailor treatments to address the underlying mechanism for hypotension in these children. (J Trauma Acute Care Surg. 2015;78: 1143Y1148. Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved.) LEVEL OF EVIDENCE: Prognostic and epidemiologic study, level III. KEY WORDS: Pediatric; hypotension; head injury. BACKGROUND:
C
urrently, it is stated by the American College of Surgeons in the Advanced Trauma Life Support (ATLS) manual, ‘‘For all practical purposes, shock does not result from isolated brain injuries.’’1 Once shock is identified in the trauma patient, resuscitation is immediately implemented, according to ATLS recommendations, with boluses of isotonic fluid, followed by blood transfusion if the shock persists. The rationale for this treatment is the assumption that most hypotension and shock in trauma are hemorrhagic, or hypovolemic, in origin. However, emerging evidence suggests that shock may occur in pediatric patients with isolated head injury.2Y4 Pediatric trauma resuscitation includes treatment in what is termed ‘‘the golden hour’’ after trauma, a time of rapid assessment and intervention during a time of peak mortality after trauma. While recognizing the progress made in injury Submitted: September 19, 2014, Revised: February 13, 2015, Accepted: February 17, 2015. From the Departments of Pediatrics and Emergency Medicine (A.R.G.), General Surgery, Division of Surgical Sciences (D.I.D.), Biostatistical Sciences (J.A.T.), Emergency Medicine (C.D.M.), and General Surgery, Section of Pediatric Surgery (J.P.), Wake Forest University School of Medicine, Winston-Salem, North Carolina. Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.jtrauma.com). Address for reprints: Alison R. Gardner, MD, MS, Department of Emergency Medicine and Pediatrics, Wake Forest School of Medicine, Medical Center Blvd., Winston-Salem, NC 27157; email: [email protected]. DOI: 10.1097/TA.0000000000000658
prevention, injury in childhood remains an incompletely preventable occurrence, and trauma remains the leading cause of death among children aged 1 to 18 years.5,6 As a step toward refining current trauma resuscitation protocols for pediatric patients with shock, our objective was to determine if children of different age groups have an increased risk of hypotension, one objective sign of shock, after head injury. Based on preliminary studies,2,4 we hypothesized that, in pediatric trauma patients presenting with hypotension after trauma, isolated head injury is associated with hypotension, particularly in younger patients. If isolated head injury proves to be mechanistically involved in the development of hypotension after trauma, it should lead us to further examine the optimal treatment strategy for these children. Pharmacologic adjuncts to fluid resuscitation could improve outcomes by decreasing the volume of fluid given, a treatment that, in the absence of blood loss, could increase cerebral edema, decrease cerebral perfusion pressure, and worsen neurologic outcomes.7 As a first step, we seek to determine whether an association exists between isolated head injury and hypotension.
PATIENTS AND METHODS Study Design Data were obtained from a national registry, the 2009 National Trauma Data Bank (NTDB). The NTDB is sponsored by the American College of Surgeons, Committee on Trauma, and
J Trauma Acute Care Surg Volume 78, Number 6
1143
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
is the largest existing database of trauma registry data. Registry data are validated by trauma registrars from participating institutions for accuracy and completeness before submission to the NTDB. This study was reviewed and approved by the Wake Forest School of Medicine Institutional Review Board.
Study Population The 2009 NTDB Report (version 7.2) is based on more than 1.3 million admission year records on more than 680,000 individuals from 567 facilities collected from the previous year.8 One hundred eighty-six of these facilities are verified as Level I adult trauma centers, representing 94% of all Level I centers, and 118 centers are verified as Level I or II pediatric trauma centers. The NTDB was queried to identify all children aged 0 to 15 years who presented to the emergency department for trauma care. Children with missing systolic blood pressure (SBP), age, sex, or who presented dead on arrival were excluded (Fig. 1).
Study Protocol and Definitions Age, sex, race, Injury Severity Score (ISS), Glasgow Coma Scale (GCS), heart rate, and initial SBP at the time of the child’s presentation to the emergency department, and discharge International Classification of Diseases, Ninth Revision (ICD-9), diagnosis codes were obtained for each child. To be consistent with the diagnosis of hypotension in clinical trauma practice and according to ATLS guidelines, children were considered to be hypotensive if the initial SBP reading on arrival to the emergency department was less than 70 mm Hg + (2 age in years) for those children younger than 10 years or less than 90 mm Hg for those children aged 10 to 15 years.
Figure 1. Cohort inclusion criteria from 2009 NTDB data. NTDB, National Trauma Data Bank; n, number; SBP, systolic blood pressure. 1144
Children were categorized into age strata including 0 to 4 years, 5 to 11 years, and 12 to 15 years chosen a priori to reflect general age levels where physiologic and anatomic changes occur and to be consistent with prior studies.2,4 The primary predictor of interest was injury pattern, classified from discharge diagnosis ICD-9 codes. ICD-9 codes were classified into injury types, with the primary type of injury of interest being isolated head injury, defined as head injury without evidence of hemorrhagic or spinal cord injury. Each ICD-9 code was classified a priori as hemorrhagic injury (indicating injury that has a known possibility of significant hemorrhage), head injury, spinal cord injury, superficial injury, other internal injury (nonhemorrhagic), facial injury, trachea/ larynx injury, other fracture (nonfemur/pelvic), late trauma complications, environmental exposures, or burns (see Appendix, Supplemental Digital Content 1, http://links.lww.com/TA/A572). The ICD-9 codes associated with each child were then used to classify the injury pattern in each child as hemorrhagic, isolated head injury, spinal cord injury, or other. Two investigators independently reviewed the classification, and any discrepancies were discussed and resolved. To obtain isolated head-injured patients as the study population of interest, we conservatively chose to classify any injuries having the possibility of causing hemorrhage as hemorrhagic injury, including scalp lacerations. In addition, and for the same purposes, any child with possible spinal cord injury was classified as a spinal cord injury type, regardless of concomitant head injury. Children with head injuries and other injury types not known to cause hypotension were classified into the isolated head injury category.
Data Analysis We compared demographic characteristics in participants who were excluded from the cohort because of missing data using W2 tests. Descriptive statistics were calculated for the entire cohort and by hypotensive status using frequencies for categoric data (age group, sex, and race) and medians and interquartile ranges for continuous variables (heart rate, ISS). Chi-square tests for categoric variables and Wilcoxon ranksum for continuous variables were used to compare the above variables by hypotension status. We used log-binomial regression to estimate relative risks (RRs). First, we examined the interaction of age group and injury pattern. Finding a significant interaction ( p G 0.0001), all further analyses were stratified by age group. Population attributable risk percent was estimated treating isolated head injury, hemorrhagic injury, and spinal cord injury as exposures.9 In sensitivity analyses, logistic regression was used to estimate the independent effects of injury pattern by age group adjusted for sex, race, and ISS. Because hypotension was rare in all subgroups, adjusted odds ratios (ORs) approximate adjusted RRs. We also estimated RR by type of head injury (cerebral hemorrhage, skull fracture, cerebral hemorrhage and skull fracture, and head injury not otherwise specified) in logbinomial models, with a 3Ydegree of freedom contrast used to compare differences by head injury type. We used the HosmerLemeshow test to evaluate the goodness of fit of the logistic regression models. Statistical analyses were performed using SAS Enterprise Guide version 4.2 (Cary, NC) and SAS (version * 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
9.3, Cary, NC). A two-sided level of > = 0.05 was used to indicate statistical significance.
RESULTS In the year 2009, the NTDB registry contained data for 682,036 patients. Using our inclusion criteria of known age of 0 to 15 years (N = 89,880), SBP measurement on arrival, and nonmissing sex and ICD-9 codes, 78,673 children were included in the analysis (Fig. 1). Of these 78,673 injured pediatric patients who constituted the final study population, 1,596 children (2.0%) were found to be hypotensive for age at presentation after trauma (Table 1). Children without SBP measures were significantly more likely to be female (39.9% of sample with missing data vs. 34.7% of sample without missing data), African American (17.9% vs. 15.4%), and in the youngest age category (68.1% vs. 31.6%). The rate of hypotension was highest in the 12- to 15-year-old children (2.3%), compared with the 0- to 4-year-old children (2.0%) and the 5- to 11-year-old children (1.8%). Hypotension rates also varied significantly by sex and race, with the highest rates of hypotension seen in females and African Americans. On average, ISS and heart rate were higher in children with hypotension. In all age groups, there were significant differences in hypotension by injury pattern (Table 2). Notably, among children with isolated head injury, the rate of hypotension was the highest in the 0- to 4-year group (2.3%). With regard to our primary objective, we found that the corresponding RR for children aged 0 to 4 years demonstrated that they were 2.5 times more likely (95% confidence interval [95% CI], 2.0Y3.2) to experience hypotension after sustaining isolated head injury than children who sustained other injuries
TABLE 2. Hypotension by Age Group and Injury Pattern* Total (N = 78,673)
Hypotensive (n = 1,596)
10,207 5,368 190 9,054
249 (2.4) 141 (2.6) 6 (3.2) 89 (1.0)
6,896 6,795 447 14,814
117 (1.7) 229 (3.4) 13 (2.9) 173 (1.2)
6,720 7,185 1,002 9,995
124 (1.9) 319 (4.4) 16 (1.6) 120 (1.2)
0Y4 y, n (%) Isolated head injury Hemorrhagic injury Spinal cord injury Other 5Y11 y, n (%) Isolated head injury Hemorrhagic injury Spinal cord injury Other 12Y15 y, n (%) Isolated head injury Hemorrhagic injury Spinal cord injury Other
p** G0.0001
G0.0001
G0.0001
*National Trauma Data Bank 2009. **p for test of injury pattern by hypotension group within age group using W2 test. N, number; %, percent.
not known to cause hypotension. There was also significantly increased RR of hypotension after isolated head injury in the older age groups; however, the excess risk associated with head injury was much lower than in the younger age group, with children in these age groups being 1.5 times more likely to experience hypotension compared with those with other injuries (Table 3). The RR for hypotension after hemorrhagic injury was significantly increased in all age groups. Furthermore, in
TABLE 1. Cohort Characteristics, Children Aged 0 to 15 Years Obtained From the NTDB 2009 Total (N = 78,673) Age, n (%), y 0Y4 5Y11 12Y15 Sex, n (%) Male Female Race, n (%) White AA/Black Other Unknown Heart rate, median (25th, 75th)** ISS, median (25th, 75th)† GCS score, median (25th, 75th)‡ AIS score (head), median (25th, 75th)§
Normotensive (n = 77,077)
Hypotensive (n = 1,596)
p* 0.0002
24,819 (31.6) 28,952 (36.8) 24,902 (31.7)
24,334 (98.0) 28,420 (98.2) 24,323 (97.7)
485 (2.0) 532 (1.8) 579 (2.3)
51,368 (65.3) 27,305 (34.7)
50,377 (98.1) 26,700 (97.8)
991 (1.9) 605 (2.2)
47,237 (60.0) 12,125 (15.4) 12,840 (16.3) 6,471 (8.2) 104 (88, 122) 5 (4, 9) 15 (15, 15) 3 (2, 4)
46,319 (98.1) 11,843 (97.7) 12,566 (97.9) 6,349 (98.1) 104 (88, 122) 5 (4, 9) 15 (15, 15) 3 (2, 4)
918 (1.9) 282 (2.3) 274 (2.1) 122 (1.9) 109 (88, 135) 10 (4, 22) 15 (6, 15) 4 (2, 4)
0.007
0.04
G0.0001 G0.0001 G0.0001 G0.0001
*p value from W2 test (age, sex, race) or Wilcoxon rank-sum test (heart rate, ISS). **n = 78,074 for heart rate. †n = 74,415 for ISS. ‡n = 73,926 for GCS. §n = 26,070 for AIS (head). NTDB, National Trauma Data Bank; n, number; %, percent; sd, standard deviation; AA, African American; ISS, Injury Severity Score; GCS, Glasgow Coma Scale; AIS, Abbreviated Injury Scale.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1145
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
TABLE 3. RR and PAR% of Hypotension by Injury Type and Age Group*
1.88Y3.30 vs. other]; cerebral hemorrhage, RR, 2.30 [95% CI, 1.73Y3.05] vs. other).
RR (95% CI)
PAR %**
DISCUSSION
Isolated head injury Hemorrhagic injury Spinal cord injury Other
2.48 (1.95, 3.16) 2.67 (2.05, 3.48) 3.21 (1.42, 7.25) 1.00
30.7 18.2 0.9 V
Isolated head injury Hemorrhagic injury Spinal cord injury Other
1.45 (1.15, 1.83)† 2.89 (2.37, 3.51) 2.49 (1.43, 4.34) 1.00
6.9 28.1 1.5 V
Isolated head injury Hemorrhagic injury Spinal cord injury Other
1.54 (1.20, 1.97)† 3.70 (3.00, 4.55) 1.33 (0.79, 2.23) 1.00
7.5 40.2 0.7 V
In this analysis, using the largest national cohort of injured pediatric patients to date, isolated head injury was significantly more likely to be associated with hypotension in younger children as compared with older children. Head injury was a significant risk factor for hypotension, with RRs of approximately the same magnitude as in hemorrhagic injury in 0- to 4-year-old children; this was not true for older children. Previous smaller studies at single institutions2,4 have demonstrated that a large proportion of younger patients presenting with hypotension had isolated head injury; however, this study was able to demonstrate that isolated head injury is a possible significant risk factor for hypotension in this age group, with a population attributable risk of 31%. The RR for isolated head injury is comparable to the RR associated with hemorrhagic injury in this age group. The RR for isolated head injury as a predictor of hypotension dramatically drops as the child’s age increases. The reason for this phenomenon remains unclear, but several possibilities exist. It is possible that the hypotension from head injury in small children could be hemorrhagic in nature given their large headYtoYbody size ratio. However, our subanalysis using types of isolated head injury found no significant differences between skull fracture and cerebral hemorrhage in predicting hypotension in the 0- to 4-year age group. Another possibility is that the hypotension in this age group is neurogenic in origin, similar to the known phenomenon of spinal shock, or caused by neurohumoral factors involved in the stress response.10,11 It is also possible that, included in the category of neurogenic shock, it can be described as an autonomic process with increased vagal tone or poor sympathetic tone.12,13 Further investigation is needed into these possible physiologic explanations. The phenomenon of hypotension after head injury is beginning to be described in pediatric patients but has also been noted in preclinical studies. In animal models, hypotension from head injury has been found to be relatively resistant to volume replacement with fluids alone, instead requiring pharmacologic interventions to correct perfusion. Fulton et al.14 showed that isolated brain injury in a swine model caused hypotension, although, again, the mechanism remains unclear. In addition, inducing head injury followed by hemorrhage in a pig model resulted in altered vascular compensation to hemorrhage and made accepted resuscitative measures with volume replacement ineffective. It remains unclear, but is highly suggested, from these early studies and animal data that head injury itself could cause the cardiovascular dysfunction of shock. No etiology or definite pathophysiology has been elucidated to explain how head injury would lead cardiovascular dysfunction and shock. Some related studies have shown general cardiovascular dysfunction in the setting of pediatric head injury. Biswas et al.15 used heart rate variability as a measure of autonomic function and demonstrated that, with traumatic brain injury and increased intracranial pressure, there was a decrease in heart rate variability in children younger than
0Y4 y
5Y11 y
12Y15 y
*National Trauma Data Bank 2009. **PAR% = population attributable risk percent. †p G 0.0001 vs. RR for hemorrhagic injury using linear contrast.
the 0- to 4-year age group, we found that the risk of hypotension with isolated head injury was not significantly different from the risk from hemorrhagic injury; in the other two age groups, the risk from hemorrhagic injury was significantly greater. To quantify the possible impact of isolated head injury on the risk of hypotension in children who experience head trauma, we calculated population attributable risk percent (Table 3). In the 0- to 4-year group, we found that almost one third of hypotension (31%) was associated with isolated head injury in contrast to about one fifth (18%) with hemorrhagic injury and only one hundredth (1%) with spinal cord injury. The population attributable risk percent for head injury was much lower in the older age groups, with values of approximately 7% for both groups. In contrast, the population attributable risk percent increased for hemorrhagic injury by age group, reaching 40% in the oldest group. In sensitivity analyses, we fit models adjusted for sex, race, and ISS; the effect of injury pattern was attenuated in all age groups. However, in the 0- to 4-year age group, those with isolated head injury remained more likely to experience hypotension (OR, 1.8; 95% CI, 1.4Y2.3 vs. other) and the risk of hypotension was not significantly different from hemorrhagic injury. In each of the other two age groups, children with isolated head injury did not experience a greater risk of hypotension (each OR, 1.0; 95% CI, 0.8Y1.3 vs. other) and the risk of hypotension with hemorrhagic injury was significantly greater (both p G 0.01). In a subanalysis of isolated head injury broken down by type (cerebral hemorrhage, skull fracture, cerebral hemorrhage and skull fracture, and head injury not otherwise specified) in 0- to 4-year-old children, we found no significant differences by type ( p = 0.32), with both skull fractures and cerebral hemorrhage ICD-9 codes significantly predicting hypotension in children aged 0 to 4 years (skull fracture, RR, 2.49 [95% CI, 1146
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
12 years. Goldstein et al.16,17 showed uncoupling of the autonomic and cardiovascular systems in acute brain injury in children younger than 18 years that was proportional to the degree of head injury. Dash et al.18 demonstrate evidence of bradycardia and electrocardiographic changes in children younger than 12 years with head injury. However, although these studies continue to suggest the ‘‘head-to-heart’’ connection, the changes in cardiovascular function as a result of head injury continue to remain unexplained, and recommendations by ATLS regarding pediatric trauma resuscitation in the face of head injury have not changed. Multiple studies have demonstrated that hypotension and hypoxia in head injury have the greatest negative impact on outcome, including increased morbidity and mortality.17,18 In fact, hypotension has been associated with significantly higher mortality rates in children as compared with adults.19 Therefore, treating hypotension in the pediatric trauma patient is of utmost importance and maybe never more so than in the headinjured child.20,21 However, some studies indicate that large amounts of fluid resuscitation with isotonic fluids, in the setting of head injury, may contribute to increased cerebral edema and worsening of secondary brain injury.7 It is possible that adjunctive treatments for hypotension after head injury in children will include an earlier addition of vasopressors, increasing sympathetic tone, or atropine to block increased vagal tone, in addition to fluid resuscitation. Pharmacologic adjuncts may decrease the amount of fluid needed to achieve a normotensive status and decrease resultant cerebral edema, increase cerebral perfusion pressure, and improve neurologic outcomes. How this will impact clinical trauma resuscitation care going forward will depend on the physiologic cause uncovered with future studies. The question of precise treatment can be argued to be most important in this young population, given the years of life left, that they may recover as fully as possible.
level of error in obtaining blood pressure measurements in children. At the same time, we recognize that hypotension is a late clinical finding in pediatric shock and poor perfusion states in children. Therefore, using hypotension as a marker of shock may have resulted in underreporting the number of children actually experiencing the tissue perfusion abnormalities of shock. In addition, as our exposure variable, we used discharge diagnosis ICD-9 codes to classify injuries as possibly hemorrhagic. This conservative classification or overclassification of some patients into the hemorrhagic injury category was made to confidently identify isolated head injury. As a result, some hypotensive children classified into the hemorrhagic injury may not have had injuries that lead to hypotension from hemorrhage, but the patient was classified as suffering from hemorrhagic injuries. Despite these limitations, our results replicate findings in our prior single-center study in which the presence of hypotension and elevated blood lactate levels were used to define shock. The same pattern was evident in both studies; children aged 0 to 4 years presenting with hypotension after trauma sustained isolated head injuries more than 50% of the time.4 We did not adjust for traditionally used as markers of injury severity, ISS and GCS, in the estimates of RR. It was felt that GCS would be affected by the presence or absence of hypotension and therefore be a mediating variable rather than a confounding variable. ISS is arguably also influenced by injury type and location; however, we did perform a sensitivity analysis including ISS in an adjusted model and our results were similar; in the 0- to 4-year age group, those with isolated head injury remained more likely to be hypotensive and the risk of hypotension was not significantly different from hemorrhagic injury.
LIMITATIONS
Hypotension occurs after isolated head injury in young children, and isolated head injury predicts hypotension as much as hemorrhagic injuries in children aged 0 to 4 years. This finding should now lead us to confirm whether a causeeffect relationship exists and, if so, isolate the responsible mechanism. In turn, this could reveal an opportunity to tailor treatments to address the underlying mechanism for hypotension in these children.
This article uses the largest national cohort of injured pediatric patients to date and as such provides power that a larger single-center study cannot. It is subject to the general limitations inherent to any database review and may not be representative of the US population. We had no way to verify correct recording of measurements and outcome variables or account for unmeasured confounding factors. However, the large sample size available in the NTDB provided us with the power to examine the predictors of hypotension, including isolated head injury. And, although we feel that the association of hypotension with head injury can logically be assumed to be one of head injury serving as a risk factor for hypotension, a true cause-and-effect relationship cannot traditionally be deduced from a cross-sectional study. The outcome variable of hypotension was defined using a singular blood pressure measurement. The initial treatment of hypotension after trauma is based, in clinical practice, on the singular blood pressure measure of a child on arrival to the trauma bay. So, in reality, our definition of hypotension after trauma is accurate with clinical practice and guidelines given by the American College of Surgeons.1 We acknowledge the
CONCLUSIONS
AUTHORSHIP J.P. and A.R.G. were responsible for the primary conception of the investigation. A.R.G. and J.A.T. had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. J.A.T. developed the analytic data set. C.D.M. and D.I.D. provided significant oversight of study design and manuscript preparation along with the significant mentorship from J.A.T. in design and statistical analysis and J.P. in clinical expertise of pediatric trauma. All authors were responsible for critical revision of the manuscript and gave final approval for submission.
DISCLOSURE A.R.G. received funding in the form of salary support from the Childress Institute for Pediatric Trauma as a Childress Scholar. This salary support
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1147
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
allowed her time release to perform the work, including the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. C.D.M. declares no conflict of interest. He would like to disclose past consultancy with Mylan Specialty LP, Bard Access Systems, Inc., and as paid expert medical malpractice case review for the Law Offices of Wade E. Byrd, PA, Lewis & Oliver; Hinshaw and Culbertson LLP. Past grants paid to his institution include those by Breathquant Medical, LLC, Johnson & Johnson/Scios Inc., EKR Therapeutics, 3 M, Commonwealth of Pennsylvania Department of Health, Dyax Corp., American College of Radiology Imaging, and Allere Scarborough, Inc. Current grants to his institution are supplied by Radiometer, Cardiorentis LTD, Zentox, and Novartis. He has received payment of lectures given on behalf of the Society of Chest Pain Centers and royalties from Up-to-Date. He has a patent filed with the US Patent and Trademark Office related to the diagnosis of chest pain. He has received research support in the form of experimental software from Siemens. D.I.D., J.A.T., and J.P. declare no conflicts of interest.
REFERENCES 1. Advanced Trauma Life Support for Doctors (ATLS). 8th ed. Chicago, IL: American College of Surgeons; 2008:55Y72. 2. Partrick DA, Bensard DD, Janik JS, Karrer FM. Is hypotension a reliable indicator of blood loss from traumatic injury in children? Am J Surg. 2002;184:555Y559. 3. Mahoney EJ, Biffl WL, Harrington DT, Cioffi WG. Isolated brain injury as a cause of hypotension in the blunt trauma patient. J Trauma. 2003;55: 1065Y1069. 4. Gardner A, Poehling KA, Miller CD, Tooze JA, Petty J. Isolated head injury is a cause of shock in pediatric trauma patients. Pediatr Emerg Care. 2013;29:879Y883. 5. Centers for Disease Control and Prevention. Vital signs: unintentional injury deaths among persons aged 0Y19 yearsVUnited States, 2000Y2009. MMWR Morb Mortal Wkly Rep. 2012;61:270Y276. 6. Kochanek KD, Kirmeyer SE, Martin JA, Strobino DM, Guyer B. Annual summary of vital statistics: 2009. Pediatrics. 2012;129:338Y348. 7. Jin G, DeMoya MA, Duggan M, et al. Traumatic brain injury and hemorrhagic shock: evaluation of different resuscitation strategies in a large animal model of combined insults. Shock. 2012;38:49Y56.
1148
8. American College of Surgeons. NTDB 2009 User Guide. Chicago, IL: American College of Surgeons; 2010. 9. Baker SP, O’Neill B, Haddon W Jr, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14:187Y196. 10. Copes WS, Champion HR, Sacco WJ, Lawnick MM, Keast SL, Bain LW. The Injury Severity Score revisited. J Trauma. 1988;28:69Y77. 11. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol. 1974;99:325Y332. 12. Masutani S, Senzaki H, Ishido H, et al. Vasopressin in the treatment of vasodilatory shock in children. Pediatr Int. 2005;47:132Y136. 13. Liedel JL, Meadow W, Nachman J, Koogler T, Kahana MD. Use of vasopressin in refractory hypotension in children with vasodilatory shock: five cases and a review of the literature. Pediatr Crit Care Med. 2002; 3:15Y18. 14. Fulton RL, Flynn WJ, Mancino M, Bowles D, Cryer HM. Brain injury causes loss of cardiovascular response to hemorrhagic shock. J Invest Surg. 1993;6:117Y131. 15. Biswas AK, Scott WA, Sommerauer JF, Luckett PM. Heart rate variability after acute traumatic brain injury in children. Crit Care Med. 2000;28: 3907Y3912. 16. Goldstein B, Kempski MH, DeKing D, et al. Autonomic control of heart rate after brain injury in children. Crit Care Med. 1996;24:234Y240. 17. Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the autonomic and cardiovascular systems in acute brain injury. Am J Physiol. 1998;275:R1287YR1292. 18. Dash M, Bithal PK, Prabhakar H, Chouhan RS, Mohanty B. ECG changes in pediatric patients with severe head injury. J Neurosurg Anesthesiol. 2003;15:270Y273. 19. Kokoska ER, Smith GS, Pittman T, Weber TR. Early hypotension worsens neurological outcome in pediatric patients with moderately severe head trauma. J Pediatr Surg. 1998;33:333Y338. 20. Pigula FA, Wald SL, Shackford SR, Vane DW. The effect of hypotension and hypoxia on children with severe head injuries. J Pediatr Surg. 1993;28: 310Y314. 21. Luerssen TG, Klauber MR, Marshall LF. Outcome from head injury related to patient’s age. A longitudinal prospective study of adult and pediatric head injury. J Neurosurg. 1988;68:409Y416.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Injury patterns associated with hypotension in pediatric trauma patients: A national trauma database review Alison R. Gardner, MD, MS, Debra I. Diz, PhD, Janet A. Tooze, PhD, MPH, Chadwick D. Miller, MD, MS, and John Petty, MD, Winston-Salem, North Carolina
Hypotension after trauma is most commonly assumed to be hemorrhagic, or hypovolemic, in origin. However, hypotension may occur in pediatric patients with isolated head injury, challenging accepted tenets of trauma care. We sought to quantify the contribution of head injury to the development of hypotension after pediatric trauma. METHODS: This is a retrospective cohort analysis using the National Trauma Data Bank registry 2009. Children aged 0 to 15 years were classified by injury pattern sustained during trauma using discharge diagnosis International Classification of Diseases, Ninth Revision, codes into isolated head, hemorrhagic, spinal cord, or other injury type. The primary outcome was hypotension for age at arrival to the emergency department. Risk of hypotension was estimated and compared by injury pattern using absolute and relative risks (RRs) stratified by age group (0Y4 years, 5Y11 years, 12Y15 years). RESULTS: Rates of hypotension ranged from 1.8% to 2.3% by age, with the highest incidence in the 12- to 15-year group. The RR of hypotension from isolated head injury (RR, 2.5; 95% confidence interval, 2.0Y3.2 vs. other) was not significantly different from the RR for hemorrhagic injury (RR, 2.7; 95% confidence interval, 2.1Y3.5 vs. other) in the 0- to 4-year-old group. For the older age groups, the RR of hypotension from isolated head injury was significantly lower than from hemorrhagic injury. CONCLUSION: Hypotension occurs after isolated head injury in children, and the risk of hypotension is as great as hemorrhagic injuries in children aged 0 to 4 years. This finding should now lead us to confirm whether a cause-effect relationship exists and, if so, isolate the responsible mechanism. In turn, this could reveal an opportunity to tailor treatments to address the underlying mechanism for hypotension in these children. (J Trauma Acute Care Surg. 2015;78: 1143Y1148. Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved.) LEVEL OF EVIDENCE: Prognostic and epidemiologic study, level III. KEY WORDS: Pediatric; hypotension; head injury. BACKGROUND:
C
urrently, it is stated by the American College of Surgeons in the Advanced Trauma Life Support (ATLS) manual, ‘‘For all practical purposes, shock does not result from isolated brain injuries.’’1 Once shock is identified in the trauma patient, resuscitation is immediately implemented, according to ATLS recommendations, with boluses of isotonic fluid, followed by blood transfusion if the shock persists. The rationale for this treatment is the assumption that most hypotension and shock in trauma are hemorrhagic, or hypovolemic, in origin. However, emerging evidence suggests that shock may occur in pediatric patients with isolated head injury.2Y4 Pediatric trauma resuscitation includes treatment in what is termed ‘‘the golden hour’’ after trauma, a time of rapid assessment and intervention during a time of peak mortality after trauma. While recognizing the progress made in injury Submitted: September 19, 2014, Revised: February 13, 2015, Accepted: February 17, 2015. From the Departments of Pediatrics and Emergency Medicine (A.R.G.), General Surgery, Division of Surgical Sciences (D.I.D.), Biostatistical Sciences (J.A.T.), Emergency Medicine (C.D.M.), and General Surgery, Section of Pediatric Surgery (J.P.), Wake Forest University School of Medicine, Winston-Salem, North Carolina. Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.jtrauma.com). Address for reprints: Alison R. Gardner, MD, MS, Department of Emergency Medicine and Pediatrics, Wake Forest School of Medicine, Medical Center Blvd., Winston-Salem, NC 27157; email: [email protected]. DOI: 10.1097/TA.0000000000000658
prevention, injury in childhood remains an incompletely preventable occurrence, and trauma remains the leading cause of death among children aged 1 to 18 years.5,6 As a step toward refining current trauma resuscitation protocols for pediatric patients with shock, our objective was to determine if children of different age groups have an increased risk of hypotension, one objective sign of shock, after head injury. Based on preliminary studies,2,4 we hypothesized that, in pediatric trauma patients presenting with hypotension after trauma, isolated head injury is associated with hypotension, particularly in younger patients. If isolated head injury proves to be mechanistically involved in the development of hypotension after trauma, it should lead us to further examine the optimal treatment strategy for these children. Pharmacologic adjuncts to fluid resuscitation could improve outcomes by decreasing the volume of fluid given, a treatment that, in the absence of blood loss, could increase cerebral edema, decrease cerebral perfusion pressure, and worsen neurologic outcomes.7 As a first step, we seek to determine whether an association exists between isolated head injury and hypotension.
PATIENTS AND METHODS Study Design Data were obtained from a national registry, the 2009 National Trauma Data Bank (NTDB). The NTDB is sponsored by the American College of Surgeons, Committee on Trauma, and
J Trauma Acute Care Surg Volume 78, Number 6
1143
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
is the largest existing database of trauma registry data. Registry data are validated by trauma registrars from participating institutions for accuracy and completeness before submission to the NTDB. This study was reviewed and approved by the Wake Forest School of Medicine Institutional Review Board.
Study Population The 2009 NTDB Report (version 7.2) is based on more than 1.3 million admission year records on more than 680,000 individuals from 567 facilities collected from the previous year.8 One hundred eighty-six of these facilities are verified as Level I adult trauma centers, representing 94% of all Level I centers, and 118 centers are verified as Level I or II pediatric trauma centers. The NTDB was queried to identify all children aged 0 to 15 years who presented to the emergency department for trauma care. Children with missing systolic blood pressure (SBP), age, sex, or who presented dead on arrival were excluded (Fig. 1).
Study Protocol and Definitions Age, sex, race, Injury Severity Score (ISS), Glasgow Coma Scale (GCS), heart rate, and initial SBP at the time of the child’s presentation to the emergency department, and discharge International Classification of Diseases, Ninth Revision (ICD-9), diagnosis codes were obtained for each child. To be consistent with the diagnosis of hypotension in clinical trauma practice and according to ATLS guidelines, children were considered to be hypotensive if the initial SBP reading on arrival to the emergency department was less than 70 mm Hg + (2 age in years) for those children younger than 10 years or less than 90 mm Hg for those children aged 10 to 15 years.
Figure 1. Cohort inclusion criteria from 2009 NTDB data. NTDB, National Trauma Data Bank; n, number; SBP, systolic blood pressure. 1144
Children were categorized into age strata including 0 to 4 years, 5 to 11 years, and 12 to 15 years chosen a priori to reflect general age levels where physiologic and anatomic changes occur and to be consistent with prior studies.2,4 The primary predictor of interest was injury pattern, classified from discharge diagnosis ICD-9 codes. ICD-9 codes were classified into injury types, with the primary type of injury of interest being isolated head injury, defined as head injury without evidence of hemorrhagic or spinal cord injury. Each ICD-9 code was classified a priori as hemorrhagic injury (indicating injury that has a known possibility of significant hemorrhage), head injury, spinal cord injury, superficial injury, other internal injury (nonhemorrhagic), facial injury, trachea/ larynx injury, other fracture (nonfemur/pelvic), late trauma complications, environmental exposures, or burns (see Appendix, Supplemental Digital Content 1, http://links.lww.com/TA/A572). The ICD-9 codes associated with each child were then used to classify the injury pattern in each child as hemorrhagic, isolated head injury, spinal cord injury, or other. Two investigators independently reviewed the classification, and any discrepancies were discussed and resolved. To obtain isolated head-injured patients as the study population of interest, we conservatively chose to classify any injuries having the possibility of causing hemorrhage as hemorrhagic injury, including scalp lacerations. In addition, and for the same purposes, any child with possible spinal cord injury was classified as a spinal cord injury type, regardless of concomitant head injury. Children with head injuries and other injury types not known to cause hypotension were classified into the isolated head injury category.
Data Analysis We compared demographic characteristics in participants who were excluded from the cohort because of missing data using W2 tests. Descriptive statistics were calculated for the entire cohort and by hypotensive status using frequencies for categoric data (age group, sex, and race) and medians and interquartile ranges for continuous variables (heart rate, ISS). Chi-square tests for categoric variables and Wilcoxon ranksum for continuous variables were used to compare the above variables by hypotension status. We used log-binomial regression to estimate relative risks (RRs). First, we examined the interaction of age group and injury pattern. Finding a significant interaction ( p G 0.0001), all further analyses were stratified by age group. Population attributable risk percent was estimated treating isolated head injury, hemorrhagic injury, and spinal cord injury as exposures.9 In sensitivity analyses, logistic regression was used to estimate the independent effects of injury pattern by age group adjusted for sex, race, and ISS. Because hypotension was rare in all subgroups, adjusted odds ratios (ORs) approximate adjusted RRs. We also estimated RR by type of head injury (cerebral hemorrhage, skull fracture, cerebral hemorrhage and skull fracture, and head injury not otherwise specified) in logbinomial models, with a 3Ydegree of freedom contrast used to compare differences by head injury type. We used the HosmerLemeshow test to evaluate the goodness of fit of the logistic regression models. Statistical analyses were performed using SAS Enterprise Guide version 4.2 (Cary, NC) and SAS (version * 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
9.3, Cary, NC). A two-sided level of > = 0.05 was used to indicate statistical significance.
RESULTS In the year 2009, the NTDB registry contained data for 682,036 patients. Using our inclusion criteria of known age of 0 to 15 years (N = 89,880), SBP measurement on arrival, and nonmissing sex and ICD-9 codes, 78,673 children were included in the analysis (Fig. 1). Of these 78,673 injured pediatric patients who constituted the final study population, 1,596 children (2.0%) were found to be hypotensive for age at presentation after trauma (Table 1). Children without SBP measures were significantly more likely to be female (39.9% of sample with missing data vs. 34.7% of sample without missing data), African American (17.9% vs. 15.4%), and in the youngest age category (68.1% vs. 31.6%). The rate of hypotension was highest in the 12- to 15-year-old children (2.3%), compared with the 0- to 4-year-old children (2.0%) and the 5- to 11-year-old children (1.8%). Hypotension rates also varied significantly by sex and race, with the highest rates of hypotension seen in females and African Americans. On average, ISS and heart rate were higher in children with hypotension. In all age groups, there were significant differences in hypotension by injury pattern (Table 2). Notably, among children with isolated head injury, the rate of hypotension was the highest in the 0- to 4-year group (2.3%). With regard to our primary objective, we found that the corresponding RR for children aged 0 to 4 years demonstrated that they were 2.5 times more likely (95% confidence interval [95% CI], 2.0Y3.2) to experience hypotension after sustaining isolated head injury than children who sustained other injuries
TABLE 2. Hypotension by Age Group and Injury Pattern* Total (N = 78,673)
Hypotensive (n = 1,596)
10,207 5,368 190 9,054
249 (2.4) 141 (2.6) 6 (3.2) 89 (1.0)
6,896 6,795 447 14,814
117 (1.7) 229 (3.4) 13 (2.9) 173 (1.2)
6,720 7,185 1,002 9,995
124 (1.9) 319 (4.4) 16 (1.6) 120 (1.2)
0Y4 y, n (%) Isolated head injury Hemorrhagic injury Spinal cord injury Other 5Y11 y, n (%) Isolated head injury Hemorrhagic injury Spinal cord injury Other 12Y15 y, n (%) Isolated head injury Hemorrhagic injury Spinal cord injury Other
p** G0.0001
G0.0001
G0.0001
*National Trauma Data Bank 2009. **p for test of injury pattern by hypotension group within age group using W2 test. N, number; %, percent.
not known to cause hypotension. There was also significantly increased RR of hypotension after isolated head injury in the older age groups; however, the excess risk associated with head injury was much lower than in the younger age group, with children in these age groups being 1.5 times more likely to experience hypotension compared with those with other injuries (Table 3). The RR for hypotension after hemorrhagic injury was significantly increased in all age groups. Furthermore, in
TABLE 1. Cohort Characteristics, Children Aged 0 to 15 Years Obtained From the NTDB 2009 Total (N = 78,673) Age, n (%), y 0Y4 5Y11 12Y15 Sex, n (%) Male Female Race, n (%) White AA/Black Other Unknown Heart rate, median (25th, 75th)** ISS, median (25th, 75th)† GCS score, median (25th, 75th)‡ AIS score (head), median (25th, 75th)§
Normotensive (n = 77,077)
Hypotensive (n = 1,596)
p* 0.0002
24,819 (31.6) 28,952 (36.8) 24,902 (31.7)
24,334 (98.0) 28,420 (98.2) 24,323 (97.7)
485 (2.0) 532 (1.8) 579 (2.3)
51,368 (65.3) 27,305 (34.7)
50,377 (98.1) 26,700 (97.8)
991 (1.9) 605 (2.2)
47,237 (60.0) 12,125 (15.4) 12,840 (16.3) 6,471 (8.2) 104 (88, 122) 5 (4, 9) 15 (15, 15) 3 (2, 4)
46,319 (98.1) 11,843 (97.7) 12,566 (97.9) 6,349 (98.1) 104 (88, 122) 5 (4, 9) 15 (15, 15) 3 (2, 4)
918 (1.9) 282 (2.3) 274 (2.1) 122 (1.9) 109 (88, 135) 10 (4, 22) 15 (6, 15) 4 (2, 4)
0.007
0.04
G0.0001 G0.0001 G0.0001 G0.0001
*p value from W2 test (age, sex, race) or Wilcoxon rank-sum test (heart rate, ISS). **n = 78,074 for heart rate. †n = 74,415 for ISS. ‡n = 73,926 for GCS. §n = 26,070 for AIS (head). NTDB, National Trauma Data Bank; n, number; %, percent; sd, standard deviation; AA, African American; ISS, Injury Severity Score; GCS, Glasgow Coma Scale; AIS, Abbreviated Injury Scale.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1145
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
TABLE 3. RR and PAR% of Hypotension by Injury Type and Age Group*
1.88Y3.30 vs. other]; cerebral hemorrhage, RR, 2.30 [95% CI, 1.73Y3.05] vs. other).
RR (95% CI)
PAR %**
DISCUSSION
Isolated head injury Hemorrhagic injury Spinal cord injury Other
2.48 (1.95, 3.16) 2.67 (2.05, 3.48) 3.21 (1.42, 7.25) 1.00
30.7 18.2 0.9 V
Isolated head injury Hemorrhagic injury Spinal cord injury Other
1.45 (1.15, 1.83)† 2.89 (2.37, 3.51) 2.49 (1.43, 4.34) 1.00
6.9 28.1 1.5 V
Isolated head injury Hemorrhagic injury Spinal cord injury Other
1.54 (1.20, 1.97)† 3.70 (3.00, 4.55) 1.33 (0.79, 2.23) 1.00
7.5 40.2 0.7 V
In this analysis, using the largest national cohort of injured pediatric patients to date, isolated head injury was significantly more likely to be associated with hypotension in younger children as compared with older children. Head injury was a significant risk factor for hypotension, with RRs of approximately the same magnitude as in hemorrhagic injury in 0- to 4-year-old children; this was not true for older children. Previous smaller studies at single institutions2,4 have demonstrated that a large proportion of younger patients presenting with hypotension had isolated head injury; however, this study was able to demonstrate that isolated head injury is a possible significant risk factor for hypotension in this age group, with a population attributable risk of 31%. The RR for isolated head injury is comparable to the RR associated with hemorrhagic injury in this age group. The RR for isolated head injury as a predictor of hypotension dramatically drops as the child’s age increases. The reason for this phenomenon remains unclear, but several possibilities exist. It is possible that the hypotension from head injury in small children could be hemorrhagic in nature given their large headYtoYbody size ratio. However, our subanalysis using types of isolated head injury found no significant differences between skull fracture and cerebral hemorrhage in predicting hypotension in the 0- to 4-year age group. Another possibility is that the hypotension in this age group is neurogenic in origin, similar to the known phenomenon of spinal shock, or caused by neurohumoral factors involved in the stress response.10,11 It is also possible that, included in the category of neurogenic shock, it can be described as an autonomic process with increased vagal tone or poor sympathetic tone.12,13 Further investigation is needed into these possible physiologic explanations. The phenomenon of hypotension after head injury is beginning to be described in pediatric patients but has also been noted in preclinical studies. In animal models, hypotension from head injury has been found to be relatively resistant to volume replacement with fluids alone, instead requiring pharmacologic interventions to correct perfusion. Fulton et al.14 showed that isolated brain injury in a swine model caused hypotension, although, again, the mechanism remains unclear. In addition, inducing head injury followed by hemorrhage in a pig model resulted in altered vascular compensation to hemorrhage and made accepted resuscitative measures with volume replacement ineffective. It remains unclear, but is highly suggested, from these early studies and animal data that head injury itself could cause the cardiovascular dysfunction of shock. No etiology or definite pathophysiology has been elucidated to explain how head injury would lead cardiovascular dysfunction and shock. Some related studies have shown general cardiovascular dysfunction in the setting of pediatric head injury. Biswas et al.15 used heart rate variability as a measure of autonomic function and demonstrated that, with traumatic brain injury and increased intracranial pressure, there was a decrease in heart rate variability in children younger than
0Y4 y
5Y11 y
12Y15 y
*National Trauma Data Bank 2009. **PAR% = population attributable risk percent. †p G 0.0001 vs. RR for hemorrhagic injury using linear contrast.
the 0- to 4-year age group, we found that the risk of hypotension with isolated head injury was not significantly different from the risk from hemorrhagic injury; in the other two age groups, the risk from hemorrhagic injury was significantly greater. To quantify the possible impact of isolated head injury on the risk of hypotension in children who experience head trauma, we calculated population attributable risk percent (Table 3). In the 0- to 4-year group, we found that almost one third of hypotension (31%) was associated with isolated head injury in contrast to about one fifth (18%) with hemorrhagic injury and only one hundredth (1%) with spinal cord injury. The population attributable risk percent for head injury was much lower in the older age groups, with values of approximately 7% for both groups. In contrast, the population attributable risk percent increased for hemorrhagic injury by age group, reaching 40% in the oldest group. In sensitivity analyses, we fit models adjusted for sex, race, and ISS; the effect of injury pattern was attenuated in all age groups. However, in the 0- to 4-year age group, those with isolated head injury remained more likely to experience hypotension (OR, 1.8; 95% CI, 1.4Y2.3 vs. other) and the risk of hypotension was not significantly different from hemorrhagic injury. In each of the other two age groups, children with isolated head injury did not experience a greater risk of hypotension (each OR, 1.0; 95% CI, 0.8Y1.3 vs. other) and the risk of hypotension with hemorrhagic injury was significantly greater (both p G 0.01). In a subanalysis of isolated head injury broken down by type (cerebral hemorrhage, skull fracture, cerebral hemorrhage and skull fracture, and head injury not otherwise specified) in 0- to 4-year-old children, we found no significant differences by type ( p = 0.32), with both skull fractures and cerebral hemorrhage ICD-9 codes significantly predicting hypotension in children aged 0 to 4 years (skull fracture, RR, 2.49 [95% CI, 1146
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
12 years. Goldstein et al.16,17 showed uncoupling of the autonomic and cardiovascular systems in acute brain injury in children younger than 18 years that was proportional to the degree of head injury. Dash et al.18 demonstrate evidence of bradycardia and electrocardiographic changes in children younger than 12 years with head injury. However, although these studies continue to suggest the ‘‘head-to-heart’’ connection, the changes in cardiovascular function as a result of head injury continue to remain unexplained, and recommendations by ATLS regarding pediatric trauma resuscitation in the face of head injury have not changed. Multiple studies have demonstrated that hypotension and hypoxia in head injury have the greatest negative impact on outcome, including increased morbidity and mortality.17,18 In fact, hypotension has been associated with significantly higher mortality rates in children as compared with adults.19 Therefore, treating hypotension in the pediatric trauma patient is of utmost importance and maybe never more so than in the headinjured child.20,21 However, some studies indicate that large amounts of fluid resuscitation with isotonic fluids, in the setting of head injury, may contribute to increased cerebral edema and worsening of secondary brain injury.7 It is possible that adjunctive treatments for hypotension after head injury in children will include an earlier addition of vasopressors, increasing sympathetic tone, or atropine to block increased vagal tone, in addition to fluid resuscitation. Pharmacologic adjuncts may decrease the amount of fluid needed to achieve a normotensive status and decrease resultant cerebral edema, increase cerebral perfusion pressure, and improve neurologic outcomes. How this will impact clinical trauma resuscitation care going forward will depend on the physiologic cause uncovered with future studies. The question of precise treatment can be argued to be most important in this young population, given the years of life left, that they may recover as fully as possible.
level of error in obtaining blood pressure measurements in children. At the same time, we recognize that hypotension is a late clinical finding in pediatric shock and poor perfusion states in children. Therefore, using hypotension as a marker of shock may have resulted in underreporting the number of children actually experiencing the tissue perfusion abnormalities of shock. In addition, as our exposure variable, we used discharge diagnosis ICD-9 codes to classify injuries as possibly hemorrhagic. This conservative classification or overclassification of some patients into the hemorrhagic injury category was made to confidently identify isolated head injury. As a result, some hypotensive children classified into the hemorrhagic injury may not have had injuries that lead to hypotension from hemorrhage, but the patient was classified as suffering from hemorrhagic injuries. Despite these limitations, our results replicate findings in our prior single-center study in which the presence of hypotension and elevated blood lactate levels were used to define shock. The same pattern was evident in both studies; children aged 0 to 4 years presenting with hypotension after trauma sustained isolated head injuries more than 50% of the time.4 We did not adjust for traditionally used as markers of injury severity, ISS and GCS, in the estimates of RR. It was felt that GCS would be affected by the presence or absence of hypotension and therefore be a mediating variable rather than a confounding variable. ISS is arguably also influenced by injury type and location; however, we did perform a sensitivity analysis including ISS in an adjusted model and our results were similar; in the 0- to 4-year age group, those with isolated head injury remained more likely to be hypotensive and the risk of hypotension was not significantly different from hemorrhagic injury.
LIMITATIONS
Hypotension occurs after isolated head injury in young children, and isolated head injury predicts hypotension as much as hemorrhagic injuries in children aged 0 to 4 years. This finding should now lead us to confirm whether a causeeffect relationship exists and, if so, isolate the responsible mechanism. In turn, this could reveal an opportunity to tailor treatments to address the underlying mechanism for hypotension in these children.
This article uses the largest national cohort of injured pediatric patients to date and as such provides power that a larger single-center study cannot. It is subject to the general limitations inherent to any database review and may not be representative of the US population. We had no way to verify correct recording of measurements and outcome variables or account for unmeasured confounding factors. However, the large sample size available in the NTDB provided us with the power to examine the predictors of hypotension, including isolated head injury. And, although we feel that the association of hypotension with head injury can logically be assumed to be one of head injury serving as a risk factor for hypotension, a true cause-and-effect relationship cannot traditionally be deduced from a cross-sectional study. The outcome variable of hypotension was defined using a singular blood pressure measurement. The initial treatment of hypotension after trauma is based, in clinical practice, on the singular blood pressure measure of a child on arrival to the trauma bay. So, in reality, our definition of hypotension after trauma is accurate with clinical practice and guidelines given by the American College of Surgeons.1 We acknowledge the
CONCLUSIONS
AUTHORSHIP J.P. and A.R.G. were responsible for the primary conception of the investigation. A.R.G. and J.A.T. had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. J.A.T. developed the analytic data set. C.D.M. and D.I.D. provided significant oversight of study design and manuscript preparation along with the significant mentorship from J.A.T. in design and statistical analysis and J.P. in clinical expertise of pediatric trauma. All authors were responsible for critical revision of the manuscript and gave final approval for submission.
DISCLOSURE A.R.G. received funding in the form of salary support from the Childress Institute for Pediatric Trauma as a Childress Scholar. This salary support
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1147
J Trauma Acute Care Surg Volume 78, Number 6
Gardner et al.
allowed her time release to perform the work, including the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. C.D.M. declares no conflict of interest. He would like to disclose past consultancy with Mylan Specialty LP, Bard Access Systems, Inc., and as paid expert medical malpractice case review for the Law Offices of Wade E. Byrd, PA, Lewis & Oliver; Hinshaw and Culbertson LLP. Past grants paid to his institution include those by Breathquant Medical, LLC, Johnson & Johnson/Scios Inc., EKR Therapeutics, 3 M, Commonwealth of Pennsylvania Department of Health, Dyax Corp., American College of Radiology Imaging, and Allere Scarborough, Inc. Current grants to his institution are supplied by Radiometer, Cardiorentis LTD, Zentox, and Novartis. He has received payment of lectures given on behalf of the Society of Chest Pain Centers and royalties from Up-to-Date. He has a patent filed with the US Patent and Trademark Office related to the diagnosis of chest pain. He has received research support in the form of experimental software from Siemens. D.I.D., J.A.T., and J.P. declare no conflicts of interest.
REFERENCES 1. Advanced Trauma Life Support for Doctors (ATLS). 8th ed. Chicago, IL: American College of Surgeons; 2008:55Y72. 2. Partrick DA, Bensard DD, Janik JS, Karrer FM. Is hypotension a reliable indicator of blood loss from traumatic injury in children? Am J Surg. 2002;184:555Y559. 3. Mahoney EJ, Biffl WL, Harrington DT, Cioffi WG. Isolated brain injury as a cause of hypotension in the blunt trauma patient. J Trauma. 2003;55: 1065Y1069. 4. Gardner A, Poehling KA, Miller CD, Tooze JA, Petty J. Isolated head injury is a cause of shock in pediatric trauma patients. Pediatr Emerg Care. 2013;29:879Y883. 5. Centers for Disease Control and Prevention. Vital signs: unintentional injury deaths among persons aged 0Y19 yearsVUnited States, 2000Y2009. MMWR Morb Mortal Wkly Rep. 2012;61:270Y276. 6. Kochanek KD, Kirmeyer SE, Martin JA, Strobino DM, Guyer B. Annual summary of vital statistics: 2009. Pediatrics. 2012;129:338Y348. 7. Jin G, DeMoya MA, Duggan M, et al. Traumatic brain injury and hemorrhagic shock: evaluation of different resuscitation strategies in a large animal model of combined insults. Shock. 2012;38:49Y56.
1148
8. American College of Surgeons. NTDB 2009 User Guide. Chicago, IL: American College of Surgeons; 2010. 9. Baker SP, O’Neill B, Haddon W Jr, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14:187Y196. 10. Copes WS, Champion HR, Sacco WJ, Lawnick MM, Keast SL, Bain LW. The Injury Severity Score revisited. J Trauma. 1988;28:69Y77. 11. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol. 1974;99:325Y332. 12. Masutani S, Senzaki H, Ishido H, et al. Vasopressin in the treatment of vasodilatory shock in children. Pediatr Int. 2005;47:132Y136. 13. Liedel JL, Meadow W, Nachman J, Koogler T, Kahana MD. Use of vasopressin in refractory hypotension in children with vasodilatory shock: five cases and a review of the literature. Pediatr Crit Care Med. 2002; 3:15Y18. 14. Fulton RL, Flynn WJ, Mancino M, Bowles D, Cryer HM. Brain injury causes loss of cardiovascular response to hemorrhagic shock. J Invest Surg. 1993;6:117Y131. 15. Biswas AK, Scott WA, Sommerauer JF, Luckett PM. Heart rate variability after acute traumatic brain injury in children. Crit Care Med. 2000;28: 3907Y3912. 16. Goldstein B, Kempski MH, DeKing D, et al. Autonomic control of heart rate after brain injury in children. Crit Care Med. 1996;24:234Y240. 17. Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the autonomic and cardiovascular systems in acute brain injury. Am J Physiol. 1998;275:R1287YR1292. 18. Dash M, Bithal PK, Prabhakar H, Chouhan RS, Mohanty B. ECG changes in pediatric patients with severe head injury. J Neurosurg Anesthesiol. 2003;15:270Y273. 19. Kokoska ER, Smith GS, Pittman T, Weber TR. Early hypotension worsens neurological outcome in pediatric patients with moderately severe head trauma. J Pediatr Surg. 1998;33:333Y338. 20. Pigula FA, Wald SL, Shackford SR, Vane DW. The effect of hypotension and hypoxia on children with severe head injuries. J Pediatr Surg. 1993;28: 310Y314. 21. Luerssen TG, Klauber MR, Marshall LF. Outcome from head injury related to patient’s age. A longitudinal prospective study of adult and pediatric head injury. J Neurosurg. 1988;68:409Y416.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
