Infection Epidemiology, Antibiotic Treatment, and Microorganism Profiles in Urban Civilian Low Energy Firearm Fractures

Abstract

Background:

Civilian fractures resulting from firearm injuries are a challenging clinical entity associated with a high risk of infection. This study aims to characterize rates, associated risk factors, and pathogen profile of infections after low-energy civilian firearm fractures.

Methods:

This study included 1,508 adults with low-velocity firearm fractures at a Level I trauma center (2018-2022). Demographic, injury, and clinical data were collected with infection status determining stratification. Outcomes of interest in the infection cohort included pathogen, timing of antibiotics, and clinical course.

Results:

Of 1,508 patients, 140 patients (9.3%) developed an infection. Factors significantly associated with infection included age, male gender, lower extremity fracture, vascular injury, compartment syndrome, visceral injury, retained foreign body, and operative intervention. Polymicrobial infections were found in 67 patients (68%), and the mean time from injury to positive culture was 70.8 ± 129.2 days.

Conclusion:

Infections after firearm fractures are a challenging complication with a predominance of methicillin-resistant Staphylococcus aureus, multidrug resistant Gram-negative organisms, and rare pathogens typically seen in immunocompromised individuals. Our study emphasizes need for further study regarding efficacy of prophylactic antibiotics and suggests utility in targeted treatment for atypical organisms.

Level of Evidence:

Level III. See Instructions for Authors for a complete description of levels of evidence.

Introduction

Firearm-related injuries in totality pose a challenge in the United States, leading to over 80,000 emergency department (ED) visits annually between 2009 and 2017^1,2. The economic burden surpasses $170 billion annually³, with complications worsening the socioeconomic impact. Rates of infection from gunshot wounds (GSWs) vary, influenced by factors including wound location, soft-tissue injury, velocity, vascular involvement, and season^4-6. In extremity low-velocity GSWs (<2000 fps), infection rates range from 1.8% to 15.7%, with conflicting evidence regarding efficacy of prophylactic antibiotics and formal irrigation and debridement^7-12. Superficial infection, deep tissue infection, and bacteremia in the setting of concomitant fracture potentiate delayed wound healing and complications including nonunion, amputation, and death¹³.

There is a paucity of literature investigating infections following civilian firearm fractures. Existing evidence, comprised of case reports, anecdotal traumatologist consensus, and 2 previous studies, suggests a pattern of atypical and antimicrobial-resistant organisms in GSWs when compared with nonpenetrating injuries^14-19. In addition, there is a lack of consensus regarding administration of prophylactic antibiotics and extent of debridement in low-velocity gunshot wounds^20,21.

The purpose of this study was to characterize the infection rate, risk factors, and microorganisms of low energy firearm fracture wounds within a large civilian cohort at a single urban trauma center. We hypothesized that infections at the site of firearm fractures will be associated with atypical (environmental/zoonotic) and drug-resistant pathogens with the potential to improve outcomes with risk-tailored antimicrobial regimens.

Materials and Methods

A retrospective review of 1,508 patients with low-energy firearm fractures at a major urban Level I trauma center between 2018 and 2022 was conducted. Patients younger than 18 years and/or with immunocompromising conditions (Appendix 1) were excluded (Fig. 1). This study evaluated low-energy injuries (minimal soft tissue loss and contamination) from low-velocity firearms as seen almost exclusively at our institution. High-energy injuries were excluded, including those injuries with large soft tissue defects or identified high energy weapon. Operative treatment included open reduction and internal fixation (ORIF), intramedullary nailing (IMN), and/or percutaneous pinning. Nonoperative treatment included immobilization in splint, cast, or sling. Per institutional practice, no patients underwent isolated irrigation and debridement for GSW injuries. The Gustilo-Anderson classification was not used for the classification of injuries as it automatically assigns GSW-related fracture to Type IIIA limiting stratification²². Patients were stratified by occurrence of bacteremia, superficial wound infection, or fracture-related infection (FRI). Superficial wound infections and FRI were combined given the retrospective nature of the study to ensure over inclusion of potential infections and avoid underdiagnosis if superficial infection patients were lost to follow-up or actually had misclassified deep infections. Cultures were taken at the discretion of the surgeon depending on the clinical findings. FRIs were determined on the basis of the 2018 consensus guidelines²³: (1) fistula, sinus, and wound breakdown; (2) purulent drainage from wound or pus during surgery; (3) pathogens identified by culture; and (4) pathogens in deep tissue confirmed by histopathology.

Fig. 1.

Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) diagram.

For the infection subgroup, data points were collected with 30-day follow-up including tobacco use, diabetes, and HbA1c. Treatment variables consisted of prophylactic antibiotics administration within 3 hours of presentation, antibiotic types, cultured pathogens, and time to infection. Outcome variables included nonunion, revision procedures, amputation, and mortality. Analysis on efficacy of prophylactic antibiotics was conducted using a propensity-matched cohort analysis. Patients were matched on a 2:1 basis using a nearest neighbor algorithm based on age, sex, and fracture location. A secondary matched analysis was performed removing polytrauma patients and factors associated with infection such as compartment syndrome, vascular injury, and visceral injury.

Statistical Analysis

Data were aggregated using Microsoft Excel (Microsoft Corporation, Redmond) and imported into R 4.5.0 (R Foundation for Statistical Computing) for analysis. Categorical variables are listed as counts with percentages and were compared using odds ratios and multivariate logistic regression. Continuous variables are expressed as means with standard deviations and were compared using the Student's t test for independent means or Wilcoxon rank sum. All statistical analyses were conducted at an α = 0.05. At an α = 0.05, β = 0.2, and early antibiotic administration (EAA) incidence of 0.6, the initial and isolated matched cohorts were powered to detect proportional differences of 0.14 and 0.27, respectively.

Ethical Statement

The study was reviewed and approved by the University of Chicago IRB.

Results

A total of 1,508 patients with low-velocity firearm fractures were included. The average age was 29.6 years (SD: 10.2, range: 18-92) with 90.4% being men. In total, 704 patients (46.7%) underwent operative fixation (ORIF—267 patients, IMN—361 patients, percutaneous pinning—76 patients). Overall, 140 patients (9.3%) developed an infection which included 20 patients (1.3%) with bacteremia, 116 patients (7.7%) with a superficial wound/fracture site infection, and 4 patients (0.3%) with both bacteremia and wound infection. Factors significantly associated with infection included age, sex, fracture location, and operative intervention (Table I). Within the infection group, 85 of 140 patients (60.7%) had undergone operative intervention (adj-OR 1.77, p = 0.002).

TABLE I.

Patient Demographics, Fracture Location, and Associated Injuries in All Patients, Infection Subgroup, and Noninfection Subgroup with Associated Odds Ratios

	Ballistic Fracture Patients
	All Patients	Infection	Noninfection
Characteristic	N = 1,508 (%)	N = 140 (%)	N = 1,368 (%)	OR (95% CI)	p
Demographics
Age (yrs), mean ± SD	29.6 ± 10.2	31.9 ± 11.7	29.3 ± 9.8		0.003
BMI, mean ± SD	25.4 ± 8.4	25.6 ± 7.2	25.4 ± 8.1		0.779
Male sex	1,363 (90.4)	136 (97.1)	1,227 (89.7)	3.91 (1.42-10.7)	0.002
Fracture location*
Hand	457	27	430	0.52 (0.34-0.81)	0.003
Shoulder and elbow	348	21	327	0.56 (0.35-0.91)	0.016
Spine	397	39	358	1.09 (0.74-1.61)	0.687
Spine + neuro deficit	251	25	226	1.10 (0.70-1.73)	0.686
Hip and pelvis	457	33	424	0.69 (0.46-1.03)	0.082
Leg	534	69	465	1.89 (1.33-2.68)	<0.001
Foot and ankle	321	35	286	1.26 (0.84-1.89)	0.278
Associated injury
Vascular injury	196 (13.0)	32 (21.9)	164 (12.0)	2.18 (1.42-3.33)	<0.001
Nerve injury	314 (20.8)	33 (23.3)	281 (20.5)	1.19 (0.79-1.80)	0.384
Compartment syndrome	91 (6.0)	25 (18.0)	66 (4.8)	4.29 (2.61-7.06)	<0.001
Visceral injury	514 (34.1)	59 (42.1)	455 (33.2)	1.46 (1.03-2.08)	0.039
Polytrauma	1,035 (68.6)	104 (74.0)	931 (68.0)	1.36 (0.91-2.01)	0.151
Retained foreign body	383 (25.4)	63 (45.0)	320 (23.4)	2.68 (1.88-3.82)	<0.001
Operative intervention	704 (46.7)	85 (60.7)	619 (45.2)	1.87 (1.31-2.67)	<0.001
Flap coverage needed	103 (6.8)	15 (10.7)	88 (6.4)	1.75 (0.98-3.11)	0.056

^*Percentages not reported in fracture location due to patient overlap from high rate of polytraumatic injury.

Bold denotes statistical significance.

Neuro deficit = incomplete or complete neurological deficit on presentation.

Fracture Location

The highest rate of infections were seen in the thigh and leg (Appendix 2). In multivariable analysis, leg fractures were significantly associated with increased infection (odds ratio [OR] 1.93, p < 0.001), while hand (OR 0.55, p = 0.008) and upper arm (OR 0.59, p = 0.035) fractures were protective (Table II). Breakdown by AO/OTA classification found in Appendix 3²⁴.

TABLE II.

Multivariable Regression Analysis of Significant Factors From Table I with Reported aORs

	Ballistic Fracture Patients
Predictive Characteristic	aOR*	95% CI	p
Hand fracture	0.55	0.36—0.90	0.008
Shoulder and elbow fracture	0.59	0.38—0.96	0.035
Leg fracture	1.93	1.35—2.77	<0.001
Vascular injury	2.28	1.47—3.54	<0.001
Visceral Injury	1.42	1.01—2.12	0.044
Compartment syndrome	4.35	2.60—7.26	<0.001
Operative intervention	1.77	1.23—2.54	0.002
Age (per year)	1.01	0.99—1.03	0.261

aOR = adjusted odds ratio.

^*Male sex and retained foreign body omitted due to near-perfect collinearity.

Bold denotes statistical significance.

Associated Injury

Concomitant injuries significantly associated with infection included vascular injury (OR 2.28, p < 0.001), compartment syndrome (OR 4.35, p < 0.001), visceral injury (OR 1.42, p = 0.044), operative intervention (OR 1.77, p = 0.002), and retained foreign body (OR 2.68, p < 0.001). No significant difference was seen between groups for concomitant nerve injury or polytraumatic injuries (Tables I and II).

Infection Subgroup

Of the 140 patients with infection, the mean age was 31.9 years (SD 11.7, range: 18-64, mean body mass index was 25.6 (SD 7.2), 136 (97.1%) were men, 16 (11.4%) were had diabetes, and 81 (57.9%) used tobacco products (Appendix 4). Ninety patients (64.3%) received antibiotics within 3 hours of presentation to the emergency department.

Matched Cohort Analysis

In the matched cohort of infection (n = 140) and noninfection (n = 280), there were no significant differences in age, sex, or fracture location (Table III). EAA was associated with a 60% reduction in the odds of infection (OR: 0.40, 95% confidence interval [CI] 0.25-0.64, p < 0.001) (Table IV). When compared with cefazolin, piperacillin-tazobactam, ceftriaxone, and clindamycin demonstrated no significant reduction in odds of infection.

TABLE III.

Patient Demographics and Fracture Location in Overall Infection and Noninfection Cohort Following Propensity Matching

	Matched Cohort—All Patients
	Infection	Noninfection
Match Characteristics	N = 140 (%)	N = 280 (%)	p
Demographics
Age (yrs), mean ± SD	31.9 ± 11.7	31.7 ± 12.1	0.872
Male Sex	136 (97.1)	272 (97.1)	1
Fracture Location
Lower Extremity	104 (74.3)	207 (73.9)	0.937
Upper Extremity	48 (34.3)	97 (34.6)	0.942
Spine	39 (27.9)	78 (27.9)	1
Pelvis	33 (23.6)	66 (23.6)	1

TABLE IV.

Early Antibiotic Administration (Within 3 Hours of Presentation) in Matched Infection and Noninfection Cohorts

	Matched Cohort Patients
	Infection	Noninfection
	N = 140 (%)	N = 280 (%)	OR (95% CI)	p
Early antibiotic administration	90 (64.3)	229 (81.8)	0.40 (0.25-0.64)	<0.001

Odds ratio reported comparing infection in patients receiving EAA vs. those without.

Bold denotes statistical significance.

In the post hoc matched analysis removing polytrauma and high-risk (concomitant vascular injury, visceral injury, compartment syndrome) patients, there were no significant differences in age, sex, or fracture location in infection patients (n = 37) and noninfection patients (n = 74) (Table V). EAA was not significantly associated with a reduction in odds of infection (OR 0.62, 95% CI 0.26-1.49, p = 0.287) (Table VI).

TABLE V.

Patient Demographics and Fracture Location in Infection and Noninfection Cohort Following Propensity Matching in Patients with Fractures without Concomitant Injuries as Below

	Matched Cohort—Fractures without Concomitant Injuries*
	Infection	Noninfection
Match Characteristics	N = 37 (%)	N = 74 (%)	p
Demographics
Age (yrs), mean ± SD	32.3 ± 9.8	31.5 ± 11.8	0.723
Male Sex	37 (100.0)	74 (100.0)	1
Fracture Location
Lower Extremity	22 (59.5)	45 (60.8)	0.891
Upper Extremity	16 (43.2)	31 (41.9)	0.892
Spine	8 (21.6)	14 (18.9)	0.763
Pelvis	6 (16.2)	12 (16.2)	1

^*Concomitant injuries include vascular injury, visceral injury, or compartment syndrome.

Bold denotes statistical significance.

TABLE VI.

Early Antibiotic Administration (Within 3 Hours of Presentation) in Matched Infection and Noninfection Cohorts in Patients without Concomitant Injuries

	Matched Cohort—Fractures without Concomitant Injuries*
	Infection	Noninfection
	N = 37 (%)	N = 74 (%)	OR (95% CI)	p
Early antibiotic administration	25 (67.6)	57 (77.0)	0.62 (0.26-1.49)	0.287

Odds ratio reported comparing infection in patients receiving EAA vs. those without.

^*Concomitant injuries include vascular injury, visceral injury, or compartment syndrome.

Bold denotes statistical significance.

Infection Diagnosis and Timeline

Of 140 patients, 120 (86% of infections, 8% overall) developed a superficial wound or fracture-related infection, and 24 (17% of infections, 1.6% of total cohort) developed bacteremia. Of 120 patients with wound infection or FRI, 74 patients (62%) were diagnosed by positive sterile deep-tissue culture and 46 (38%) were diagnosed clinically. Forty-seven percent of infections were polymicrobial. The mean time to positive culture from injury was 70.8 ± 129.2 days (bacterial infections—78.8 ± 176.3 days; fungal infections—25.1 ± 34.1 days). The mean time to negative culture and absence of composite clinical concern, measured from initial positive culture, was 24.3 ± 156.9 days (Appendix 4). There was no significant correlation between infection rate and weather measured by mean monthly local temperature (r = 0.17, p = 0.240) (Fig. 2).

Fig. 2.

Infection rate and local temperature variation over time.

Outcomes

In the infection cohort, the mean follow-up was 180.5 days (SD 327.4); 57 patients (40.1%) developed a nonunion, 22 patients (15.7%) underwent an amputation, and 19 patients (13.6%) died from any cause within follow-up. Five patients (3.6%) died during hospitalization from systemic inflammatory response. Ninety-four patients (67.1%) underwent a revision procedure including irrigation and debridement and implant removal or exchange for nonunion.

Pathogen Profile and Antimicrobial Treatment

Seventy-five genotypically distinct pathogens were isolated from 98 patients including 37 (49.3%) Gram-positive bacteria, 25 (33.3%) Gram-negative bacteria, and 13 (17.3%) fungi (Fig. 3). Methicillin-resistant Staphylococcus aureus was isolated in 22 patients (22.4%) with a mean time to growth from presentation of 163.2 days. Of the fungal infections, the most common genus was Candida, with 13 distinct pathogens of Candida albicans isolated in 14 (14.3%) patients. A breakdown of pathogen, time to growth, culture site, and antimicrobial regimen is found in Appendices 5 and 6.

Fig. 3.

Pathogen breakdown in infection cohort.

Discussion

The aims of this study were to (1) investigate the risk of fracture-related infections and bacteremia after firearm fractures, (2) identify risk factors, and (3) profile pathogens in civilian firearm fracture infections. Our aim in including all patients despite heterogeneity was to understand the role of GSWs and associated fracture-risk factors in the general population. Our findings revealed an infection rate of 9.3%, with a significantly elevated risk associated with lower extremity fractures (thigh and leg), compartment syndrome, vascular injury, retained foreign bodies, and operative intervention. Notably, the infection cohort exhibited a high rate (68%) of polymicrobial infections, with methicillin-resistant S. aureus (MRSA) and C. albicans being the predominant bacteria and fungi, respectively.

The lone study characterizing firearm fracture infections by Ghali et al. demonstrated an infection rate of 8.3% in a cohort of 347 patients suggesting similar findings of higher infection rates in lower extremity fractures¹⁸. These findings concur with general nonfirearm literature on fracture-related infections. In a meta-analysis of 6,425 fractures, Kortram et al. found that lower extremity fractures had a significantly elevated risk (RR = 1.94, p < 0.001) of developing an infection than upper extremity fractures. Hines et al. found a similar infection rate of 16.7% in a cohort of 370 patients with nonfirearm injuries undergoing emergent fasciotomy for acute compartment syndrome²⁵. Our results concerning the associated risk of vascular injury are likely attributable to the critical and well-established importance of neovascularization and tissue perfusion in wound healing and infection eradication^26,27.

In our cohort, initial operative intervention was associated with a higher risk of infection. These results may be innate to surgery or secondary to selection bias given higher severity injuries necessitating operative management and only patients undergoing procedures sampled for sterile cultures. Although our study did not stratify by fracture pattern, Ghali et al. found that the zone of comminution was significantly greater in patients with infection (63.94 mm vs. 48.46 mm, p = 0.032)¹⁸. In addition, in the face of conflicting evidence on the utility of early prophylactic antibiotics^9,20,21, our results demonstrate a significant reduction in odds of infection in the overall matched cohort, but not in patients with isolated orthopaedic injuries. However, further study is needed given the isolated orthopaedic injury matched cohort analysis being adequately powered for >27% difference.

Our study identified diverse pathogens and time to positive culture (range: 6-354 days) with a notable frequency of polymicrobial infections, aligning with existing literature associating such infections with penetrating and polytrauma²⁸. Sudduth et al. reported a 51.1% polymicrobial infection rate in 451 patients with open fractures²⁹. Alarmingly, like Sudduth et al., we observed a marked rise in the proportion of MRSA. Contextually, 36% of Staphylococcus aureus samples from our emergency department (2022-2023) were methicillin resistant, and city-wide MRSA colonization prevalence in intensive care units (2008-2013) was 11.1%.³⁰

We identified 75 bacteria and fungi, including environmental opportunistic organisms that infect the elderly or immunocompromised^31,32. Stenotrophomonas maltophilia stands out, recognized by the World Health Organization as an urgent threat due to its growing prevalence and antibiotic resistance³³. It is typically found in patients with malignancy, cystic fibrosis, or undergoing immunosuppression³⁴. In addition, we isolated Citrobacter species—C. koserii and C. freundii, affecting mainly neonates and the elderly^35,36.

Notable fungi include mucormycosis-related Lichthemia corymbifera and Rhizopus genus, typically seen in pediatric, elderly, and immunocompromised patients^37,38. Fusarium solani, a soil fungus with rising infection rates and limited treatment options, was also found³⁹. These findings highlight importance of fungal cultures in firearm fractures and concern for increased susceptibility in young, otherwise healthy individuals.

Finally, we showed no correlation between season and infection rate. Previous studies in patients with sepsis and battlefield wound infections demonstrated seasonal variation in infection microbiology^6,40. In 198 patients with battlefield injuries, Soderstrom et al. found a significant difference in proportion of Gram-negative and anaerobic bacteria isolated based on season with no difference in Gram-positive bacteria. Direct comparison of battlefield data to our study is limited by the difference in scope of injuries; however, military data remain a primary source of dedicated GSW-injury literature.

These data are relevant considering the lack of consensus guidelines on antibiotic prophylaxis or utility of debridement in firearm fractures^20,21,41. Informed guidelines are salient in the wake of large trials demonstrating local vancomycin and tobramycin powders reduce infection rates in high-risk fracture care while not affecting pathogen resistance patterns and being economically justifiable^42-45.

This study has several limitations. As a retrospective study, it is subject to bias of results and interpretation. Furthermore, all patients were seen at a single trauma center limiting the external validity. Furthermore, variable patient follow-up and inadequate power limit conclusions drawn from outcomes and EAA isolated cohort analysis, respectively. In this study, the mean follow-up was 180 days, and the mean time to positive cultures was 70 days. This timeline in the setting of our average time to infection ought to capture most potential infectious complications. Confounding factors include overestimation of infections given inclusion of superficial wound infections, institutional changes to antibiotic, culture, and treatment protocols. Although routine culturing is not performed at our institution, subjective differences in clinical concern may lead to false positives. While further study including operative fixation type and timing in a prospective and randomized manner is optimal, feasibility is limited in this patient population.

Our data indicate factors that raise infection likelihood and question the efficacy of prophylactic antibiotics in isolated low-risk fractures. Furthermore, these data suggest consideration of expanded coverage for MRSA and multidrug resistant Gram-negative bacteria in high-risk firearm fractures. Given inherent challenges of conducting prospective studies and the relationship between infection and functional outcomes⁴⁶, this study represents the best available evidence with strong potential to impact management and patient outcomes.

Appendix

Supporting material provided by the authors is posted with the online version of this article as a data supplement at jbjs.org (http://links.lww.com/JBJSOA/B24). This content was not copyedited or verified by JBJS.