Infection and Antibiotic Agents in Bleeding Trauma Patients: A Review of Available Literature

Abstract

Background:

Antibiotic prophylaxis is a common, established practice at trauma centers worldwide for patients presenting with various forms of serious injury. Many patients simultaneously present with hemorrhage. The current guidelines by the Eastern Association for the Surgery of Trauma recommend re-dosing prophylactic antibiotic agents for every 10 units of blood products administered. However, these guidelines are only mildly supported by dated research.

Methods:

A literature search was completed through Medline EBSCO Host using antibiotic prophylaxis and transfusion as keywords. Articles judged to be relevant to the study question were selected for full-text review. Case studies were not included. Altogether, 18 articles were cited in our results through this process.

Results:

Risk of infection increases in patients resuscitated with large volume of blood products. Animal models of trauma offered conflicting findings on whether blood loss and blood resuscitation altered tissue antibiotic concentrations compared with controls. Studies focused on antibiotic pharmacokinetics in non-trauma human patients revealed agreement surrounding reported decreases in serum and tissue concentrations, although there was discrepancy surrounding the clinical relevancy of the reported decreases.

Conclusions:

Trauma, hemorrhage, and transfusion impair the immune response resulting in increased incidence of infection. Both animal and human models of antibiotic pharmacokinetics show decreased serum and tissue concentrations during hemorrhage. However, available data are insufficient to conclude that trauma patients experiencing hemorrhage are at elevated risk of infection and thus require more frequent redosing of antibiotic agents than the current guidelines suggest. An upcoming, prospective study by our institution seeks to evaluate this question.

Multi-system organ failure from sepsis is the leading cause of death among hospitalized trauma patients surviving beyond 48 hours and is therefore a substantial burden to the trauma population [1]. The prophylactic use of antibiotic agents in trauma patients is an established practice at trauma centers worldwide aimed at reducing surgical site infections, abdominal abscess, empyema, and bacteremia among this particularly vulnerable population. Although local practice varies, most centers agree that prophylactic antibiotic agents are indicated in patients presenting with open fractures, penetrating abdominal trauma, and pneumothorax or hemothorax requiring surgical intervention [2]. A distinguishing factor among trauma patients is that a substantial percentage of patients present simultaneously in hemorrhagic shock. Currently, there are no adequate guidelines recommending the appropriate dosing and redosing of antibiotic agents in patients who are actively bleeding, requiring blood transfusion, and meeting indication for antibiotic agents. Additionally, in recent years, resuscitation has emphasized balanced component therapy to improve outcomes after resuscitation that even further limits the relevancy of previous research in the field of antibiotic guidance of a patient with a substantial transfusion requirement [3].

In this article, we review the pertinent literature describing the relation between hemorrhage, resuscitation, and subsequent infection among badly injured trauma patients in a stepwise fashion as follows: the potential contributing factors and mechanism of infection in trauma; infection risk among trauma patients with substantial transfusion requirements; basic science research on the effect of traumatic hemorrhage and resuscitation on antibiotic pharmacokinetics; and the effect of large volume blood loss among elective surgical patients on the pharmacokinetics of antibiotics at both the serum and tissue level. Of note, there are no studies prospectively evaluating antibiotic pharmacokinetics among trauma patients; therefore, the current evidence regarding the appropriate dosing of antibiotic agents for trauma patients during resuscitation in hemorrhagic shock is limited, thus emphasizing the need for ongoing research in this field.

Methods

A literature search was conducted in Medline EBSCO Host using antibiotic prophylaxis and transfusion as words in subject heading (MW) criteria (Fig. 1). The search identified 78 articles, all of which were subject to abstract review. Eight articles were included for full-text review, of which two were referenced in the results of this study. Entries rejected for full-text review consisted namely of specific case studies and studies not relevant to our study question. All of the references contained in the eight articles used for full-text review were themselves subjected to abstract review. Through this process, 30 additional articles underwent full-text review, of which 16 were referenced in the results of this study. Altogether, 18 articles were referenced in the results of this study and an additional three studies were referenced in the background and one in the discussion for a total of 22 citations (Table 1).

FIG. 1.

Literature search and review diagram. Medline EBSCO Host was used to search, review, and reference the first round of articles (following black arrows). Seventy-eight articles were found through the search. All 78 articles underwent abstract review; eight articles were selected for full-text review. The references for these eight articles were subjected to abstract review; 30 references were selected for full-text review (following gray arrows). Of the 38 articles subjected to full-text review, 18 articles are referenced in our results.

Table 1.

Literature Reviewed by Topic

Article Subsection	Authors	Year published	Prospective vs. retrospective	Sample size	Model
The immunosuppression of traumatic hemorrhagic shock	Davis et al.	1983	Prospective	9	Animal
	Stephan et al.	1987	Prospective	30	Animal
	Livingston et al.	1988	Prospective	75	Animal
Risk of infection increases with quantity of blood transfused among trauma patients	Dellinger et al.	1984	Prospective, retrospective	126, 212	Human
	Nichols et al.	1984	Prospective	145	Human
	Dawes et al.	1986	Retrospective	137	Human
	Ericsson et al.	1989	Prospective	150	Human
	Kirton et al.	2000	Prospective	317	Human
	Goldberg et al.^a	2012	n/a	n/a	n/a
Antibiotic pharmacokinetics among animal models of trauma	Dickson et al.	1987	Prospective	10	Animal
	Brothers et al.	1991	Prospective	30	Animal
	McKindley et al.	1995	Prospective	10	Animal
Decreased serum antibiotic concentrations in patients with hemorrhage	Levy et al.	1990	Prospective	14	Human
	Dupon et al.	1993	Prospective	10	Human
	Swoboda	1996	Prospective	11	Human
	Markantonis et al.	2004	Prospective	16	Human
Decreased tissue antibiotic concentrations in patients with hemorrhage	Swoboda	1996	Prospective	11	Human
	Dehne et al.	2001	Prospective	40	Human
	Dotters-Katz et al.	2019	Prospective	20	Human

Referenced literature has been organized by article subheading and primary authors within the first and second columns. Publication year, prospective vs. retrospective study, sample size, and study model are included in subsequent columns.

n/a = not available.

The most recent prophylactic antibiotic redosing guidelines published by the Eastern Association for the Surgery of Trauma.

Results

Immunosuppression of traumatic hemorrhagic shock

The trauma patient population is hypothesized to be uniquely prone to infection likely as a result of several factors including decreased tissue perfusion in the setting of hemorrhagic shock, mechanical contamination of the wound with foreign material at the time of injury, and direct contamination from hollow viscus injury. However, the exact mechanism by which infection risk is increased in trauma patients in hemorrhagic shock is unknown. A 1983 large animal study by Davis et al. [4] identified an impaired neutrophil migratory response in the setting of hemorrhage. Additionally, this study described a defect in neutrophil adherence in the presence of epinephrine and corticosteroids, which are both elevated in response to hemorrhage. Two later studies found that not only hemorrhage, but blood transfusion itself, led to immunosuppression in animal models. A 1987 study by Stephan et al. [5] found decreased T cell blastogenesis in response to reinfusion of autologous blood in a mouse model, and one year later, Livingston et al. [6] found a significant increase in susceptibility to infection among hemorrhaging mice. In this study, mice were subjected to hemorrhage and then randomly assigned to receive blood product resuscitation or no blood product resuscitation. The mice were then inoculated with specific quantities of Staphylococcus aureus, ranging from 10⁶ to 10¹⁰ colony forming units (CFU), and treated with various regimens of antibiotic agents. Regardless of antibiotic regimen, resuscitated mice exhibited an increased infection rate that correlated with bacteria dosage. These studies suggest that in addition to the inherent risk of the trauma population to infection, both bleeding and blood product transfusion further increase the risk of infection and sepsis among our patients.

Risk of infection increases with quantity of blood transfused among trauma patients

The current guidelines published by the Eastern Association for the Surgery of Trauma (EAST) suggest that among trauma patients experiencing blood loss, the administered dose of antibiotic agents may be “increased two to threefold and repeated after transfusion of every 10 units of blood until there is no further blood loss” [7]. These guidelines cite a 1989 prospective study from Ericsson et al. [8] that reported an increased risk of infection in patients experiencing “high blood loss,” described as “≥6 units blood loss.” In this study, infection rates were compared among trauma patients requiring exploratory laparotomy and undergoing one of three prophylactic antibiotic regimens. Each patient's blood loss was estimated based on the amount of blood units needed to raise the patients’ hemoglobin level to within normal range within the first 24 hours. Additionally, a 1984 study by Nichols et al. [9] referenced an increased infection risk for every 10 blood unit increments transfused. The EAST guidelines based upon these studies from the 1980s are considered level 3 evidence only.

Also in the 1980s, Dellinger et al. [10] completed a comprehensive prospective observational study focusing on postoperative infection risk in patients who were treated with exploratory laparotomy. Dellinger and his team reported that 3.2 units of transfused blood increased the risk of infection incrementally. The study further distinguished that the increased risk of infection was not correlated with the degree of hypotension but instead with transfusion requirement. Depending on the quantity of blood units transfused, the infection rate among transfused patients ranged anywhere from double (6–10 units) to sextuple (≥15 units) the rate of patients not requiring transfusion. In a 1986 retrospective analysis by Dawes et al. [11], similar findings were reported for patients presenting with colon injury. Although only 7.5% of patients not receiving transfusion therapy developed infectious complications, the percentage of patients receiving transfusion therapy who developed infection ranged from 25% (quadruple) for patients receiving one to five units of blood products to 57% (nearly octuple) for patients receiving ≥10 units of blood products.

These findings agree with more recent assessments of elevated infection risk in surgical patients that report an increased risk of infection after transfusion of four units of blood products. In a 2000 prospective study involving patients presenting with penetrating abdominal injury and hollow viscus injury, Kirton et al. [12] found blood transfusions greater than or equal to four units to be independently associated with surgical site infection (Fig. 2).

FIG. 2.

A visual timeline of studies finding specific resuscitative blood volumes which correlated to increased infection. Five studies referenced in this review described specific blood volumes that correlated to an increased risk of infection. This figure illustrates each of these articles plotted by the specific volume of blood products referenced versus year published. *A significantly increased risk of infection was noted between patients receiving 0 units and 1–5 units (average, 3) of blood products. **The most recent prophylactic antibiotic re-dosing guidelines published by the Eastern Association for the Surgery of Trauma.

Antibiotic pharmacokinetics among animal models of trauma

Multiple early large animal studies have evaluated antibiotic serum and tissue concentrations in the setting of hemorrhage, trauma, or both. A 1987 study by Dickson et al. [13] using dogs found that cefazolin and gentamicin serum levels did not vary substantially between hemorrhaged samples and non-hemorrhaged controls. The study evaluated 10 dogs; each dog served as its own control. For the control study, both cefazolin and gentamicin were administered to the dogs. Blood samples were drawn from zero hours to six hours and then analyzed for cefazolin and gentamicin serum concentrations. One week later, hemorrhage was induced by bleeding to a mean arterial pressure of 50 mm Hg, maintained for one hour. The dogs were then subjected to blood resuscitation using the entire volume of shed blood plus 1 L of 0.9% sodium chloride (5 dogs) or fluid resuscitation using 0.9% sodium chloride equal to three times the volume of shed blood (5 dogs). Twenty minutes after resuscitation, the dogs were infused with cefazolin and gentamicin and blood samples were obtained and analyzed as described above. Although hemorrhage was not correlated to changes in blood serum concentrations, the study noted that pharmacokinetic changes were correlated to each kind of resuscitation. Significant findings included a 30% increase in cefazolin half-life in dogs resuscitated with whole blood, a reduced clearance of cefazolin in dogs resuscitated with saline, and a 38% increase in gentamicin half-life in dogs resuscitated with saline.

In 1991, Brothers et al. [14] evaluated tissue cefazolin clearance in rabbits. Thirty rabbits experienced either no hemorrhage (10 rabbits), hemorrhage equal to 50% of blood volume (10 rabbits), or hemorrhage equal to 100% of blood volume (10 rabbits). Cefazolin was administered at 30 mg/kg. Cefazolin levels in retroperitoneal fat samples were measured for all rabbits from 30 minutes to 3.5 hours post-infusion; the same was done in iliac artery samples from 1.5 to 3.5 hours post-infusion. The study found that at 1.5 hours post-infusion, hemorrhage reduced the tissue concentration of cefazolin in both fat and vascular tissue samples between non-bled controls and animals bled 100% of their blood volume. No changes in tissue concentrations were noted between animals bled 50% and 100% of their blood volume, nor at any other time between the 100% bled and control groups.

In 1995, McKindley et al. [15] evaluated the pharmacokinetics of vancomycin and aztreonam in both trauma and hemorrhagic shock in 10 pigs. As in the study by Dickson et al. [13], the pigs served as their own controls by undergoing vancomycin and aztreonam infusion and six hours of blood sampling without induction of trauma and hemorrhage. Three days later, trauma was induced to the hind quarter, producing soft-tissue hematoma. Hemorrhage was then induced through bleeding of 40% of total blood volume (mean, 1.9 L). This shocked state was maintained for one hour. Resuscitation followed via administration of shed blood plus two times the shed volume as lactated Ringer solution. Thirty minutes later, aztreonam and vancomycin were infused, and blood samples were drawn as previously described. Both antibiotic agents were re-dosed and blood samples re-drawn on days four and eight after the same procedure. The study found multiple changes in pharmacokinetic parameters after the induction of hemorrhage and trauma. The central volume of aztreonam was decreased by 29% from day three to day eight whereas its steady state volume decreased by 34% from day one to day eight. For vancomycin, the central volume was reduced between 14% and 22% on days four and eight, respectively, compared with day three, whereas the steady state volume decreased 30% from day one to day eight. No changes were noted in the half-life of aztreonam, but the half-life of vancomycin was measured to be 17% to 52% lower than the control at all data points after induction of hemorrhage and trauma.

Together, these studies suggest that pharmacokinetic differences in the settings of trauma and hemorrhage are drug specific. Some drugs, such as cefazolin, may be more resilient to pharmacokinetic changes when administered during trauma or hemorrhage. Further research exploring the pharmacokinetics of popular prophylactic antibiotic agents in hemorrhaging and injured patients is needed.

Decreased antibiotic serum concentration in non-trauma patients with high blood loss

With the understanding that bleeding trauma patients receiving blood transfusion are at an incrementally increasing risk of developing infection, it is imperative to understand the effect of bleeding and transfusion on antibiotic pharmacokinetics in the blood. A 1993 study by Dupon et al. [16] including liver transplant patients found that trough concentrations of antibiotic agents varied widely during hemorrhage. This study found that a prophylactic regimen of 4 g piperacillin followed by a 2 g re-dose every four hours would be inadequate to maintain minimum inhibitory concentration (MIC) in the blood serum, potentially because of high blood loss and transfusion requirements among these patients.

Also in the 1990s, a prospective study by Swoboda [17] investigated the effects of blood loss on serum concentrations of prophylactic antibiotic agents. Eleven patients undergoing spinal surgery were included and received one of two antibiotic prophylaxis regimens, either cefazolin and gentamicin [10] or gentamicin and vancomycin [1]. Standard prophylactic doses of gentamicin and cefazolin were evaluated. Patients received one intravenous dose of peri-operative antibiotic agents. The average administration time before incision was two minutes for patients treated with gentamicin and 17 minutes for patients treated with cefazolin. The average blood loss was approximately 2 L whereas the average fluid replacement volume, including blood products, colloid, and crystalloid was approximately 5 L. The author referenced 4 mcg/mL and 16 mcg/mL as the cefazolin MICs for Staphylococcus aureus and Escherichia coli, respectively, and 4 mcg/mL as the gentamicin MIC for both gram-positive and gram-negative organisms. The study found that cefazolin serum concentrations were non-linearly related to blood loss, whereas gentamicin concentrations were loosely, but not substantially, correlated with blood loss. Furthermore, four of 10 (40%) patients treated with cefazolin were measured to have serum concentrations below the MIC for Escherichia coli but above the MIC for Staphylococcus aureus whereas seven of 11 (64%) patients treated with gentamicin were measured to have serum concentrations below the gram- positive/-negative MIC.

A similar 1990 prospective study by Levy et al. [18] including craniomaxillofacial surgery patients found that mean antibiotic serum concentrations were lower in surgical patients compared with non-surgical controls. The study compared the pharmacokinetics of prophylactic cloxacillin in 16 patients undergoing face and neck surgery to non-surgical controls. The mean blood loss during surgery was approximately 2 L and the mean volume of packed red blood cells administered during surgery was approximately 1 L. The quantity of cloxacillin in circulation was estimated through area under the curve (AUC) calculations of serum cloxacillin concentrations over time. The study found that mean systemic exposure to cloxacillin (AUC0→∞) was 71% lower in the operative patients compared to the non-surgical controls. This was attributed to loss of antibiotic during hemorrhage, especially within the first hour of antibiotic administration because serum antibiotic concentrations preceding systemic distribution are higher. Furthermore, three of 14 surgical patients (21%) were recorded to have insufficient cloxacillin serum MICs at four hours after dosage; this increased to eight of 14 patients (57%) by six hours, suggesting the inadequacy of the cloxacillin prophylaxis standards.

Again, a non-trauma study by Markantonis et al. [19] including 16 colorectal surgery patients found that gentamicin serum and tissue concentrations correlate with volume of fluid replacement. The average total fluid replacement approached 7 L, consisting of 71% crystalloid, 14% blood products, 11% colloid, and 4% antibiotic solution. The single prophylactic dose (2 mg/kg) of gentamicin administered was found to be insufficient to maintain MIC in blood serum and tissue in a substantial number of patients [19].

Although the studies reviewed here noting an overall insufficient serum concentration of antibiotic agents in bleeding patients do not involve a trauma population, this information may be extrapolated to patients who have suffered injuries and may therefore be at an even higher risk of infection because of direct contamination and stress response. Assuming the trauma population may undergo standard antibiotic dosing with little regard for antibiotic loss to the environment or the dilutional effects of resuscitation, the treatment team may infer that like these elective surgical patients, our patients may be receiving insufficient antibiotic agents.

Decreased antibiotic tissue concentration in non-trauma patients with high blood loss

Further studies have sought to investigate not only the antibiotic pharmacokinetics at the serum level in bleeding patients but instead at the level at which antibiotic agents are effective in the tissue. In the previously cited study by Swoboda [17], tissue concentrations of cefazolin and gentamicin were also evaluated. Among the patients treated with prophylactic cefazolin, one of 10 (10%) patients was recorded to have a tissue concentration less than 4 mcg/mL. Although cefazolin tissue concentrations rarely dropped below the MIC, they were directly related to blood loss, with changes becoming significant 60 minutes after incision. The study attributed this difference to cefazolin's high affinity for protein, meaning that at any given time, 70% to 90% of cefazolin is protein bound, leaving only 10% to 30% available for antibiosis. Among patients treated with prophylactic gentamicin, nine of 11 (82%) were measured to have tissue concentrations of gentamicin below 4 mc/mL [17].

In a 2001 prospective study, Dehne et al. [20] sought to evaluate the pharmacokinetics of various antibiotic regimens in the setting of hemorrhage including 40 patients undergoing major lower extremity orthopedic surgery. The study evaluated four different prophylactic antibiotic regimens: cefuroxime 1.5 g, cefotiam 2 g, cefamandole 2 g, and ampicillin-subulactam 2 g/1 g, all administered within 10 minutes of anesthesia induction. The mean duration of surgery ranged from 226 minutes to 235 minutes for each antibiotic regimen. The average hemorrhage observed in the study ranged from 770 mL to 880 mL (approximately 3–4 units of blood). The authors found that peri-operative antibiotic agents must be re-dosed after four hours to ensure sufficient antimicrobial bone concentration.

A 2019 study by Dotters-Katz et al. [21] noted findings different from those described above. In a prospective study of 20 pregnancies undergoing caesarean delivery, the study found no difference in tissue concentrations of prophylactic cefazolin between patients of the lower quantitative blood loss (QBL) quartile (median, 0.6 L) and higher QBL quartile (median, 1.2 L) [21]. It is important to note that this study did not compare these cefazolin tissue levels to any non-hemorrhaged control patients, so although the tissue levels between both hemorrhaging groups may be similar, it is impossible to infer from these data whether hemorrhage itself does or does not affect cefazolin tissue levels in patients undergoing caesarean delivery. Together, these results question whether high blood loss is accompanied by decreased tissue concentration of prophylactic antibiotic, potentially rendering the prophylaxis regimen ineffective.

Discussion

Trauma patients experience a high rate of infection, and the studies reviewed here suggest that the rate of infection is substantially higher in patients who also experience hemorrhage and blood product transfusion. Despite this knowledge, there is little information available on the ideal initial antibiotic dosage as well as the frequency at which hemorrhaging trauma patients should be re-dosed with prophylactic antibiotic agents. The available literature on infection rate among hemorrhaging trauma patients is limited, and although the studies agree that hemorrhage increases an injured patient's risk of infection, they diverge on the quantity of hemorrhage that results in increased infection risk.

Further highlighting the impact of antibiotic loss in hemorrhaging patients, a 2020 study by our institution using an ex vivo model of the Cell Saver Elite system found that 97% of administered antibiotic was collected in the waste bag and never returned to the patient. This study demonstrated that only 3% of the infused antibiotic was found in the re-infusion bag, implying that this commonly utilized resuscitation strategy among trauma patients may return only 3% of antibiotic agents to patients via autologous transfusion [22]. Combined with the studies cited herein that have pointed to decreased antibiotic serum/tissue concentrations and increased risk of infection both during hemorrhage, this recent finding raises further concerns about the effectiveness of current antibiotic prophylaxis practices.

Additionally, there is no available literature studying antibiotic pharmacokinetics at the serum or tissue level of trauma patients. Research on this subject using large animal models is rare, dated, and inconsistent. We are therefore left to review the insufficient literature in the colorectal, transplant, orthopedic, and obstetrical populations that implies that the standard antibiotic dosing regimens for bleeding patients receiving volume may be insufficient to maintain adequate antibiotic efficacy. Furthermore, the nomenclature of lost blood volume, sometimes defined as “blood loss” and other times as “blood products administered,” varies from study to study reducing the applicability of these findings.

To try and reduce the burden of infection among our trauma patients, our own urban, level 1 trauma center has recently established an antibiotic redosing increment to coincide with every four units of blood products. For simplicity throughout resuscitation, one “unit” of blood product can include one unit of packed red blood cells (pRBC), one unit of fresh frozen plasma (FFP), or one unit of whole blood. For example, if a patient has received two units of whole blood (which is commonly administered pre-hospital or in our trauma bay) and subsequently receives one unit of pRBC and one FFP, this would be an indication to re-dose antibiotic agents.

This literature review provides the background for our upcoming prospective study measuring the serum and tissue pharmacokinetics of antibiotics in bleeding trauma patients using remnant (leftover) blood samples and tissue. We hope that this important research will provide the much-needed data to establish guidelines for the dosing and redosing of antibiotic agents in bleeding patients with the hope of reducing the burden of infection and sepsis among trauma patients.

Footnotes

Authors' Contributions

Fabio Saccomanno is the primary author. Jonathan Gates and Lenworth Jacobs equally contributed to offering their relevant expertise in the fields of trauma surgery and infection. Joseph Kuti helped to interpret the relevance and accuracy of the studies using his expertise in pharmacy and antibiotic pharmacokinetics. Daniel Ricaurte dedicated substantial time to literature review and revision of multiple drafts of the manuscript. Jane Keating supervised Fabio and directed his work, mentoring him throughout the process by assisting in the review of literature and revising and editing all drafts of the manuscript.

Funding Information

No funding was received for this project.

Author Disclosure Statement

The authors declare no conflicts of interest.

References

Baker

, et al. Epidemiology of trauma deaths. Am J Surg, 1980;July [Epub ahead of print; DOI: 10.1016/0002-9610(80)90431-6].

Smith

, Fox

, Fakhro

, et al. “SCIP”ping antibiotic prophylaxis guidelines in trauma: The consequences of noncompliance. J Trauma Acute Care Surg, 2012; 73:452–456.

Holcomb

, Tilley

, Baraniuk

, et al. Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: the PROPPR randomized clinical trial. JAMA, 2015; 313:471–482.

Davis

, Stevens

, Peitzman

, et al. Neutrophil migratory activity in severe hemorrhagic shock. Circulatory Shock, 1983; 10:199–204.

Stephan

, Kupper

, Geha

, et al. Hemorrhage without tissue trauma produces immunosuppression and enhances susceptibility to sepsis. Arch Surg, 1987; 122:62–68.

Livingston

, Shumate

, Polk

, Malangoni

. More is better. Antibiotic management after hemorrhagic shock. Am J Med, 1988; 451–457.

Goldberg

, Anand

, Como

, et al. Prophylactic antibiotic use in penetrating abdominal trauma: An Eastern Association for the Surgery of Trauma practice management guideline. J Trauma Acute Care Surg. 2012; 73(5 Suppl 4):321–325.

Ericsson

, Fischer

, Rowlands

, et al. Prophylactic antibiotics in trauma: The hazards of underdosing. J Trauma, 1989; 29:1356–1361.

Nichols

RoL

, Smith

, Klein

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

10.

Dellinger

, Oreskovich

, Wertz

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

11.

Dawes

, Aprahamian

, Condon

, Malangoni

. The risk of infection after colon injury. Surgery, 1986; 100:796–803.

12.

Kirton

, O'Neill

, Kestner

, Tortella

. Perioperative antibiotic use in high-risk penetrating hollow viscus injury: A prospective randomized, double-blind, placebo-control trial of 24 hours versus 5 days. J Trauma, 2000; 49:822–832.

13.

Dickson

, DiPiro

, Michael

, et al. Effect of hemorrhagic shock on cefazolin and gentamicin pharmacokinetics in dogs. Antimicrob Agents Chemother. 1987; 31:389–392.

14.

Brothers

, Wakefield

, Schaberg

, et al. Effect of controlled hemorrhage on tissue and serum cefazolin clearance. J Surg Res, 1991; 51:223–228.

15.

Mckindley

, Fabian

, Boucher

, et al. Antibiotic pharmacokinetics following fluid resuscitation from traumatic shock. Arch Surg, 1995; 130:1321–1329.

16.

Dupon

, Janvier

, Vinçon

, et al. Plasma levels of piperacillin and vancomycin used as prophylaxis in liver transplant patients. Eur J Clin Pharmacol, 1993; 45:529–534.

17.

Swoboda

. Does intraoperative blood loss affect antibiotic serum and tissue concentrations?. Arch Surg, 1996; 131:1165.

18.

Levy

, Egersegi

, Strong

, et al. Pharmacokinetic analysis of cloxacillin loss in children undergoing major surgery with massive bleeding. Antimicrob Agents Chemother, 1990; 34:1150–1153.

19.

Markantonis

, Kostopanagiotou

, Panidis

, et al. Effects of blood loss and fluid volume replacement on serum and tissue gentamicin concentrations during colorectal surgery. Clin Ther, 2004; 26:271–281.

20.

Dehne

, Mühling

, Sablotzki

, et al. Pharmacokinetics of antiobiotica prophylaxis in major orthopedic surgery and blood-saving techniques. Orthopedics, 2001; 24:665–669.

21.

Dotters-Katz

, Smid

, Grace

, et al. The effect of blood loss on cefazolin levels in women undergoing cesarean delivery. Am J Perinatol, 2019; 36:688–694.

22.

Lasko

, Conelius

, Serrano

, et al. Impact of intraoperative cell salvage on concentrations of antibiotics used for surgical prophylaxis. Antimicrob Agents Chemother, 2020; 64:1389–1395.