Mechanism of Injury May Influence Infection Risk from Early Blood Transfusion

Abstract

Background:

Blood transfusion is a known risk factor for infection in trauma patients. Differences based on mechanism of injury have not been well described. We hypothesize that infection risk in trauma patients with early blood transfusion is different based on blunt or penetrating mechanism of injury.

Patients and Methods:

Adults admitted to the trauma intensive care unit from January 2010 through January 2015 were reviewed retrospectively. Those receiving transfusion after 24 h were excluded. Infections were defined as positive bronchoalveolar lavage, blood, urine, wound, or abdominal cultures. Logistic regression identified independent predictors of infection. Significance was considered at p ≤ 0.05.

Results:

With blunt trauma (n = 625), the transfusion rate was 36% (n = 223), with 30% (n = 186) infections. Those with an infection were more severely injured, had a higher operation rate, lower Glasgow Coma Score (GCS), longer hospital stay, and higher transfusion rate (all p < 0.001). With penetrating trauma (n = 292), the transfusion rate was 54% (n = 159), with 26% (n = 77) infections. Those with an infection were older, more severely injured, had a higher operation rate, lower GCS, longer length of stay, and higher transfusion rate (all p < 0.01). Controlling for age, injury severity score (ISS), revised trauma score (RTS), GCS, and hospital stay, transfusion was an independent predictor of infection in patients with blunt (odds ratio: 2.1, 95% confidence intervals: 1.272–3.393, p = 0.003) but not penetrating trauma.

Conclusions:

Early blood transfusion increases infection risk in blunt but not penetrating trauma.

Trauma patients are classified broadly by either blunt or penetrating injury mechanism, but Advanced Trauma Life Support guidelines do not distinguish care or management based on the mechanism of injury [1]. However, we [2 –5] and others [6 –11] have reported multiple outcome differences in matched blunt and penetrating injuries even with similar resuscitation end points. From these existing data, we can speculate that mechanism of injury may influence a physiologic compensatory response.

In 1946, the immune altering properties of blood transfusions were first recognized by the Nobel laureate, Sir Peter Medawar [12]. In the subsequent 70 years, there have been hundreds of studies showing that transfusion potentiates infection risk in virtually every type of surgical patient [13 –17], including trauma [18 –23]. Surprisingly, there are little data on potential differences between blunt and penetrating trauma with respect to infection risk after blood transfusion.

In 1992, Agarwal et al. [18] showed that units of blood transfused were an independent predictor of infection in both blunt and penetrating trauma. Because there have been several advancements in the trauma management and guidelines since this time, we sought to re-evaluate this topic. We hypothesized that the infection risk after early transfusion of blood may be different based on mechanism of injury.

Patients and Methods

This was a retrospective cohort study conducted at the Ryder Trauma Center in the University of Miami/Jackson Memorial Medical Center and was approved by the Institutional Review Board with waiver of informed consent. All adult blunt or penetrating trauma patients admitted to the intensive care unit (ICU) at Ryder Trauma Center from January 2010 through January 2015. Exclusions were those less than 18 years old, incarcerated, pregnant, first transfusion more than 24 h, and/or deaths less than 72 h. Early transfusion was defined as receiving blood products within 24 h of admission. Infection was defined as bronchoalveolar lavage greater than 10⁴ colony forming units (CFU), or positive blood, urine, wound, or abdominal/thoracic cultures.

All statistical analyses were performed using SPSS Statistics, version 22.0 (IBM Corporation, Armonk, NY). Categorical variables are expressed as frequency (percent) and compared between groups using χ² or Fisher exact test, as appropriate. Parametric data are expressed as mean ± standard deviation and compared using Student t test for two independent samples. Non-parametric data are expressed as median (interquartile range) and compared using the Mann-Whitney U test. Variables demonstrating statistical significance on univariable analyses were entered into a multivariable logistic regression model. Odds ratios (OR) and 95% confidence intervals (CI) are reported. All results were considered statistically significant at p ≤ 0.05.

Results

The study population comprised 917 patients, 625 with blunt trauma and 292 with penetrating trauma. In the blunt trauma group, 36% (n = 223) received transfusions and 30% developed an infection (n = 186). In the penetrating trauma group, 54% (n = 159) received transfusions and 26% (n = 77) developed an infection. These groups are compared in Table 1.

Table 1.

Baseline Characteristics by Mechanism of Injury

	Blunt (n = 625)	Penetrating (n = 292)	p
Age, years	47 ± 19	35 ± 15	<0.001
Male gender, % (n)	71.2% (579)	84.6% (274)	<0.001
Heart rate, beats/minute	96 ± 23	97 ± 23	NS
Systolic blood pressure, mm Hg	134 ± 32	117 ± 35	<0.001
Glasgow Coma Score	13 ± 4	14 ± 3	<0.001
Base deficit	−2 ± 5	−5 ± 6	<0.001
Hematocrit, %	39 ± 7	37 ± 7	<0.001
Hospital length of stay, days	14 (7–30)	11 (6–21)	0.001
Injury severity score	22 ± 12	18 ± 12	<0.001
Revised trauma score	8 (7–8)	8 (7–8)	NS
Mortality, % (n)	0.9% (7)	1.2% (4)	NS
Operation >2 h, % (n)	18.1% (147)	52.8% (171)	<0.001
Transfused, % (n)	50.6% (411)	59.0% (191)	0.012
Any infection, % (n)	33.8% (275)	27.5% (89)	0.041
Gram negative	50.9% (140)	47.2% (42)	NS
Gram positive	21.1% (58)	14.6% (13)	NS
Mixed, gram negative/positive	28.0% (77)	38.2% (34)	NS
Positive blood culture, % (n)	11.9% (97)	13.3% (43)	NS
Positive urine culture, % (n)	15.1% (123)	10.5% (34)	0.045
Positive BAL, % (n)	26.4% (214)	15.2% (49)	<0.001
Positive wound culture, % (n)	2.0% (16)	3.4% (11)	NS
Positive abscess culture, % (n)	2.2% (18)	4.3% (14)	NS

BAL = bronchoalveolar lavage; NS = not significant.

Infection incidence was plotted against packed red blood cell (PRBC) units for both blunt and penetrating trauma (Fig. 1). Both groups appear to show dose dependency, with a more linear relation in the blunt trauma patients (p < 0.001). To evaluate whether transfusion was truly an independent predictor of infection, univariable analysis and multivariable logistic regression was used.

FIG. 1.

Dose dependency curves. Infection incidence is plotted against packed red blood cell (PRBC) units for both blunt and penetrating trauma. Correlation coefficients (R²) are shown for each trend.

In the blunt trauma group, those who developed an infection were more severely injured, had a lower admission Glasgow Coma Scale score (GCS), more prolonged operations, higher Abbreviated Injury Score (AIS) Head, and a longer length of stay (LOS) compared with those who had no infection (all p < 0.001). Additionally, 63% of those developing an infection received transfusions, compared with 24% of those not receiving a transfusion (p < 0.0001). These results are summarized in Table 2.

Table 2.

Blunt Trauma Characteristics

	Negative infection (n = 439)	Positive infection (n = 186)	p
Male gender, % (n)	71.3% (313)	75.8% (141)	NS
Age, years	47 ± 20	46 ± 18	NS
Injury severity score	18 ± 10	26 ± 12	<0.001
Revised trauma score	8 (8–8)	7 (6–8)	<0.001
Glasgow Coma Score	15 (14–15)	13 (6–15)	<0.001
Operation >2 h % (n)	12.8% (56)	28.0% (52)	<0.001
AIS >2 h, % (n)	26.4% (142)	55.3% (152)	<0.001
Hospital length of stay, days	8 (5–14)	34 (22–59)	<0.001
Transfused, % (n)	24.1% (106)	62.9% (117)	<0.001

AIS = Abbreviated Injury Score; NS = not significant.

Controlling for injury severity score (ISS), revised trauma score (RTS), GCS, operation more than two hours, AIS Head >2, LOS, and transfusion status, independent predictors of infection after blunt trauma were LOS (odds ratio [OR]: 1.047, 95% confidence interval [CI]: 1.034–1.060, p < 0.001), blood transfusion (OR: 1.841, 95% CI: 1.167–2.904, p = 0.009), and AIS Head >2 (OR: 1.823, 95% CI: 1.194–2.782, p = 0.005). Increased GCS had a protective effect (OR: 0.905, 95% CI: 0.836–0.980, p = 0.015).

In the penetrating trauma group, those with infection were older, more severely injured, had a lower admission GCS, more prolonged operations, and longer LOS (all p ≤ 0.022). The transfusion rate in those developing an infection was 79%, compared with 46% in those without an infection (p < 0.0001). These results are summarized in Table 3.

Table 3.

Penetrating Trauma Characteristics

	Negative infection (n = 215)	Positive infection (n = 77)	p
Male gender, % (n)	84.2% (181)	90.9% (70)	NS
Age, years	33 ± 14	38 ± 15	0.022
Injury severity score	15 ± 10	24 ± 15	<0.0001
Revised trauma score	8 (7–8)	8 (6–8)	0.002
Glasgow Coma Score	14 ± 2	12 ± 4	<0.0001
Operation >2 h, % (n)	48.4% (104)	66.2% (51)	0.007
Hospital length of stay, days	8 (4–13)	30 (18–65)	<0.0001
Transfused, % (n)	45.6% (98)	79.2% (61)	<0.0001

NS = not significant.

Controlling for age, ISS, RTS, GCS, operation longer than two hours, LOS, and transfusion status, the only independent predictor of infection after penetrating trauma was LOS (OR: 1.100, 95% CI: 1.069–1.132). Transfusion was not a significant independent predictor of infection in this group.

Discussion

The results of this study confirm that LOS is a major risk factor for infection [24,25]. The major new finding is that blood transfusion increased the risk of infection after blunt but not penetrating trauma.

In 1978, Opelz et al. [26] noticed that kidney transplant recipients who were transfused pre-transplant had better allograft survival than those who did not receive transfusions. In 1993, Agarwal et al. [18] showed that infection incidence was increased with transfusion after both blunt and penetrating trauma. Others have described a higher incidence of infection in trauma patients after transfusion, but have not compared blunt versus penetrating injuries [19,20,22,27,28]. In addition, the age of the stored PRBCs probably influences infection [29].

We saw trends similar to Agarwal et al.'s [18] dose dependency of PRBCs versus infection risk. It follows logic that the more blood that is transfused, the more alterations there will be in the immune response. Whereas blunt and penetrating trauma groups show a linear correlation, there is a stronger and more linear relation in the blunt trauma group (p < 0.001).

These results showed that risk of infection after transfusion depended on the blunt or penetrating mechanism of injury, which seemingly contradicts the findings of Agarwal et al. [18]. However, exact comparisons between the studies are confounded by the indications, relative volumes, timing, product ratios, and end points of transfusions. Agarwal et al. [18] analyzed data on 5,366 consecutive patients hospitalized for more than two days at eight hospitals over a two-year period. We analyzed data on 917 patients over a four-year period at one hospital. Agarwal et al. [18] reported an incidence of infection of 8.9% after penetrating injuries and 12.9% after blunt injuries with a mean ISS of 12, whereas we reported an incidence of 27.5% and 33.8%, respectively, with a mean ISS of 20. The most likely reasons for the difference are that our patients were considerably sicker and that we defined infection as either a positive blood, urine, bronchoalveolar lavage, or abscess culture whereas Agarwal et al. [18] defined infection by ICD-9 coding.

In the 23 years since the study by Agarwal et al. [18], there have been several innovations in critical care, including shorter antibiotic duration, use of extracorporeal membrane oxygenation, and different ventilation strategies and fluid management. Despite this, critical care of the trauma patient remains similar regardless of the blunt or penetrating mechanism of injury. Thus, we cannot likely attribute a difference in infection to contrasting management practices. Whereas our blunt trauma patients were more severely injured, according to ISS, we still find that transfusion independently predicts infection in blunt, but not penetrating, trauma, when ISS is controlled. To facilitate comparison with previous studies, we controlled for the same factors: age, ISS, RTS, GCS, LOS, and transfusion status. However, there are differences based on mechanism of injury.

In the penetrating trauma group, univariable analysis still shows a difference in transfusion status between patients with infection and no infection. Despite this, transfusion status falls out of the multivariable logistic regression. This is likely the result of contributing risk factors. By controlling for these factors, transfusion does not have the same effect on infection risk as it does after blunt trauma. The only remaining factor predicting infection in this group was LOS, which can be somewhat controversial. There are two ways this interacts with infection rate. First, infection rate can increase because of longer LOS, and second, LOS can increase because of the treatment of an infection. It is likely that both scenarios are present in some way in all cases, making it unknown which is the cause and which is the effect.

These results should be considered in context with a series of recent studies that have shown a profound effect of injury mechanism on multiple outcomes. For example, even when injuries are matched carefully for demographics, severity, and vital signs, there are blunt versus penetrating differences in the risks of mortality and acute kidney injury after fluid resuscitation [2 –5], the risk of venous thromboembolism [30,31], and the stress-induced hyperglycemic response [30,31].

It is beyond the scope of the present study to provide a definitive explanation for the difference between blunt and penetrating trauma. However, Moore et al. [32] published work supporting the influence of such differences being driven by the effects of shock-induced hypoperfusion. Animal models have shown that neurohumoral mediators and/or cytokines released by tissue injury may play a role also [33,34], but variable combinations of tissue injury and shock make this difficult to assess in humans. Blunt trauma tends to have elements of shock with a more diffuse spread of tissue injury, compared with penetrating trauma that has shock with a relatively localized area of tissue injury. This might explain why the compensatory physiologic response to blunt versus penetrating trauma may be different, and thus result in different infection rates. This idea can be supported by two studies done by Rowell et al. [35,36], showing differences based on injury mechanism in response to massive transfusion and mortality at high ISS.

There are several limitations to this study. First, it is a retrospective review at a single institution. We are only provided with the information that is attainable from the medical record. Variables such as open body cavities and tracheostomy use are not always recorded reliably in the medical record, and using such variables could be inconsistent and confounding. Additionally, age of the blood products used was not collected because it is not made available in the medical record. Prospective data would best confirm these results, as the patient could be followed closely to record possible confounding factors. Additionally, positive infections were assessed by retrospective review of culture data. It is theoretically possible that our data collection may have missed infections because of a lack of cultures. Furthermore, type II error is possible because the penetrating trauma group in much smaller than the blunt trauma group.

In conclusion, transfusion within 24 h predicts infection independently in blunt but not penetrating trauma. This supports accumulating evidence that the mechanism of injury might influence the pathophysiologic compensatory response during recovery.

Footnotes

Acknowledgments

We thank Dr. Gerardo Guarch, as well as Dr. Krishnamurti Rao for their technical assistance in clinical data collection. We also thank Ronald J. Manning ARNP, MSPH, Clinical Research Coordinator at the Ryder Trauma Center, for his assistance in regulatory and compliance-related matters.

C.A.K. was responsible for all aspects of the study; collection, statistical analysis, and interpretation of data; drafting and revision of manuscript, figures, and tables; J.P.M., J.J.R., D.H., X.D.R., collection and interpretation of data; revision of manuscript, figures, and tables; C.I.S., N.N., interpretation of data, critical review, and revision of the manuscript;

K.G.P., overall responsibility for the study, including conception, experimental design, data analysis and interpretation, revision of manuscript, figures, and tables; supervision.

Supported in part by grants from the Naval Medical Research Center, and the U.S. Army Medical Research and Material Command

Author Disclosure Statement

No competing financial interests exist.

References

Kortbeek

, Al Turki

, Ali

, et al. Advanced trauma life support, 8th edition, the evidence for change. J Trauma, 2008; 64:1638–1650.

Allen

, Valle

, Jouria

JM, AS

, et al. Differences between blunt and penetrating trauma after resuscitation with hydroxyethyl starch. J Trauma Acute Care Surg, 2014; 77:859–864.

Ryan

, Ogilvie

, Pereira

, et al. Effect of hetastarch bolus in trauma patients requiring emergency surgery. J Spec Oper Med, 2012; 12:57–67.

Allen

, Ruiz

, Meizoso

, et al. Is hydroxyethyl starch safe in penetrating trauma patients?. Mil Med, 2016; 181(5 Suppl):152–155.

Ogilvie

, Pereira

, McKenney

, et al. First report on safety and efficacy of hetastarch solution for initial fluid resuscitation at a level 1 trauma center. J Am Coll Surg, 2010; 210:870–880.

Martin

, Mullenix

, Steele

, et al. Functional outcome after blunt and penetrating carotid artery injuries: Analysis of the National Trauma Data Bank. J Trauma, 2005; 59:860–864.

George

, McGwin

Jr , Windham

, et al. Age-related gender differential in outcome after blunt or penetrating trauma. Shock, 2003; 19:28–32.

Schreiber

, Meier

, Tisherman

, et al. A controlled resuscitation strategy is feasible and safe in hypotensive trauma patients: Results of a prospective randomized pilot trial. J Trauma Acute Care Surg, 2015; 78:687–695.

Holcomb

. Fluid resuscitation in modern combat casualty care: Lessons learned from Somalia. J Trauma, 2003; 54(5 Suppl):S46–51.

10.

McSwain

, Champion

, Fabian

, et al. State of the art of fluid resuscitation 2010: Prehospital and immediate transition to the hospital. J Trauma, 2011; 70(5 Suppl):S2–10.

11.

James

, Michell

, Joubert

, et al. Resuscitation with hydroxyethyl starch improves renal function and lactate clearance in penetrating trauma in a randomized controlled study: The FIRST trial (Fluids in Resuscitation of Severe Trauma). Br J Anaesth, 2011; 107:693–702.

12.

Medawar

. Relationship between the antigens of blood and skin. Nature, 1946; 157:161.

13.

Murphy

, Connery

, Hicks

Jr ., Blumberg

. Homologous blood transfusion as a risk factor for postoperative infection after coronary artery bypass graft operations. J Thorac Cardiovasc Surg, 1992; 104:1092–1099.

14.

Leal-Noval

, Rincon-Ferrari

, Garcia-Curiel

, et al. Transfusion of blood components and postoperative infection in patients undergoing cardiac surgery. Chest, 2001; 119:1461–1468.

15.

Banbury

, Brizzio

, Rajeswaran

, et al. Transfusion increases the risk of postoperative infection after cardiovascular surgery. J Am Coll Surg, 2006; 202:131–138.

16.

Tartter

. Blood transfusion and infectious complications following colorectal cancer surgery. Br J Surg, 1988; 75:789–792.

17.

Houbiers

, van de Velde

, van de Watering

, et al. Transfusion of red cells is associated with increased incidence of bacterial infection after colorectal surgery: A prospective study. Transfusion, 1997; 37:126–134.

18.

Agarwal

, Murphy

, Cayten

, Stahl

. Blood transfusion increases the risk of infection after trauma. Arch Surg, 1993; 128:171–176.

19.

Dellinger

, Oreskovich

, Wertz

, et al. Risk of infection following laparotomy for penetrating abdominal injury. Arch Surg, 1984; 119:20–27.

20.

Edna

, Bjerkeset

. Association between blood transfusion and infection in injured patients. J Trauma, 1992; 33:659–661.

21.

Hill

, Frawley

, Griffith

, et al. Allogeneic blood transfusion increases the risk of postoperative bacterial infection: A meta-analysis. J Trauma, 2003; 54:908–914.

22.

Rosemurgy

, Hart

, Murphy

, et al. Infection after injury: Association with blood transfusion. Am Surg, 1992; 58:104–107.

23.

Silliman

, Moore

, Johnson

, et al. Transfusion of the injured patient: proceed with caution. Shock, 2004; 21:291–299.

24.

Olaechea

, Ulibarrena

, Alvarez-Lerma

, et al. Factors related to hospital stay among patients with nosocomial infection acquired in the intensive care unit. Infect Control Hosp Epidemiol, 2003; 24:207–213.

25.

Glance

, Stone

, Mukamel

, Dick

. Increases in mortality, length of stay, and cost associated with hospital-acquired infections in trauma patients. Arch Surg, 2011; 146:794–801.

26.

Opelz

, Terasaki

. Improvement of kidney-graft survival with increased numbers of blood transfusions. N Engl J Med, 1978; 299:799–803.

27.

Nichols

, Smith

, Klein

, et al. Risk of infection after penetrating abdominal trauma. N Engl J Med, 1984; 311:1065–1070.

28.

Nichols

, Smith

. Risk of infection, infecting flora and treatment considerations in penetrating abdominal trauma. Surg Gynecol Obstet, 1993; 177(suppl):50–54.

29.

Offner

, Moore

, Biffl

, et al. Increased rate of infection associated with transfusion of old blood after severe injury. Arch Surg, 2002; 137:711–716.

30.

Karcutskie

, Meizoso

, Ray

, et al. Mechanism of injury may alter risk for venous thromboembolism after trauma. JAMA Surg, 2016;PMID: 27682749.

31.

Meizoso

, Ray

, Karcutskie

, et al. Admission hyperglycemia is associated with different outcomes after blunt versus penetrating trauma patients. J Surg Res, 2016; 206:83–89.

32.

Moore

, Moore

, Liras

, et al. Acute fibrinolysis shutdown after injury occurs frequently and increases mortality: A multicenter evaluation of 2,540 severely injured patients. J Am Coll Surg, 2016; 222:347–355.

33.

Brohi

, Cohen

, Ganter

, et al. Acute coagulopathy of trauma: Hypoperfusion induces systemic anticoagulation and hyperfibrinolysis. J Trauma, 2008; 64:1211–1217.

34.

Chesebro

, Rahn

, Carles

, et al. Increase in activated protein C mediates acute traumatic coagulopathy in mice. Shock, 2009; 32:659–665.

35.

Rowell

, Barbosa

, Diggs

, et al. Effect of high product ratio massive transfusion on mortality in blunt and penetrating trauma patients. J Trauma, 2011; 71(2 Suppl 3):S353–357.

36.

Rowell

, Barbosa

, Diggs

, et al. Specific abbreviated injury scale values are responsible for the underestimation of mortality in penetrating trauma patients by the injury severity score. J Trauma, 2011; 71(2 Suppl 3):S384–388.