Gunshot Injuries of the Hand: Incidence, Treatment Characteristics, and Factors Associated With Healthcare Utilization

Abstract

Background

The objectives of this study are to: (1) describe the demographics, injury patterns, and treatment characteristics of patients who sustained a gunshot injury (GSI) of the hand; and (2) examine the utilization of healthcare resources in patients with a GSI of the hand.

Methods

We retrospectively identified 148 adult patients who were treated for a GSI of the hand between January 2000 to December 2017 using multiple International Classification of Diseases Ninth and Tenth Edition (ICD-9 and ICD-10) codes. We used bivariate and multivariable analysis to identify which factors are associated with unplanned reoperation, length of hospitalization, and number of operations.

Results

Multivariable logistic regression showed that fracture severity was associated with unplanned reoperation. Multivariable linear regression showed that fracture severity is associated with a higher number of hand operations after a GSI of the hand, and that a retained bullet (fragment) and patients having gunshot injuries in other regions than the hand had a longer length of hospitalization. Seventy (47%) patients had sensory or motor symptoms in the hand after their GSI, of which 22 (15%) patients had a transection of the nerve.

Conclusions

Sensory and motor nerve deficits are common after a GSI of the hand. However, only 31% of patients with symptoms had a transection of the nerve. A retained bullet (fragment), having more severe hand fractures, and GSI in other regions than the hand are associated with a higher number of operation and a longer period of hospitalization.

Introduction

Firearm injuries inflict a significant burden on the U.S. healthcare system each year. Among industrialized nations, the United States has one of the highest incidences of firearm-related deaths each year with the number continuing to grow.^1,2 The incidence of injuries to the upper extremity varies from 22% to 32.2% of all gunshot injuries (GSIs).^3-5

The high energy of ballistic injuries causes significant bony and soft tissue damage surrounding the entrance and exit locations. Hand injuries caused by firearms frequently involve nerve, tendon, and/or vessel injury in addition to phalangeal and metacarpal bone fractures. ³ Although the majority of GSI of the hand are managed nonoperatively, some patients undergo one or more operations for reconstruction of the injured hand. ⁶

With an estimated annual cost of $2.1 billion, nonfatal GSIs place a significant burden on the healthcare system. ⁷ Prior studies investigating GSIs of the hand have not fully examined the extent of surgical and hospital resources consumed following these injuries, or which factors are associated with utilization of these resources. ⁶

Given the impact that GSIs of the hand may place on healthcare providers and the healthcare system, we aim to present our experience at a large academic tertiary center over a 17-year period. The objectives of this study are to: (1) describe the demographics, injury patterns, and treatment characteristics of patients who sustained GSIs of the hand; and (2) examine the utilization of healthcare resources in patients with a GSI of the hand.

Materials and Methods

After approval from our institutional review board, we retrospectively identified patients who were treated for a GSI between January 2000 to December 2017 using multiple International Classification of Diseases Ninth and Tenth Edition (ICD-9 and ICD-10) codes (Supplemental Material 1). Inclusion criteria included patients who suffered a GSI to the hand, defined as distal to the wrist crease. After manual chart review, we excluded patients who did not have GSI, patients with injuries from air operated gun types such as a ball bearing gun or pellet gun, patients whose GSI did not involve the hand, patients who were younger than 18 years of age, patients with incomplete records, and patients who died within 30 days after admission. Our final cohort consisted of 148 adult patients who were treated for a GSI of the hand (Supplemental Figure 1).

Medical charts were manually reviewed to collect data regarding patient, injury, and treatment characteristics along with indications for reoperation. Severity of the GSI was ranked based on the presence of hand fracture (phalangeal, metacarpal, and/or carpal fractures) with or without bone loss. To classify fracture severity, we used a modified version of the Orthopedic Trauma Association classification for bone loss in open fractures in which the category “no bone loss” is divided into “no fracture” and “fracture without bone loss.” This established the following classification: (1) no fracture; (2) fracture without bone loss; (3) fracture with partial bone loss (bone missing or devascularized bone fragment, but still some contact between proximal and distal fragment); and (4) fracture with segmental bone loss. Nerve deficit was defined as sensory and/or motor symptoms at presentation thought to be due to either contusion, partial transection, or complete transection. Intraoperative findings describing nerve deficits were recorded after review of operative records and were classified as the following: (1) intact/contusion; (2) partial transection; and (3) complete transection. If the operative notes did not describe any detail about nerve injury, we assumed that there was no nerve exploration and that the nerve was in continuity. Surgical treatments for nerve injury at the time of exploration were categorized as the following: (1) direct nerve repair; (2) nerve repair with autograft; (3) nerve repair with allograft; (4) neurolysis alone; and (5) no nerve intervention. Occupation was categorized in 7 groups: unemployed, retired, laborer (work dominated by lifting and grasping and repetitive hand motion), clerical work, skilled worker (including salesman, entrepreneur, and musician), law enforcement or security, and students.

Unplanned reoperation was defined as a reoperation to treat unexpected complication including failure of fixation, nonunion or delayed union, stiffness, unsalvageable digit, or failed revascularization. A planned reoperation was defined as a surgical treatment that was planned to be conducted in 2 or more stages, such as removal of an external fixation device or bone grafting after multiple debridements in the case of infection.

Health care utilization was evaluated based on the following factors: length of hospital stay, the number of operations, the number of post-op visits, and the number of radiographic studies undergone. The length of initial hospital stay was calculated based on the number of days the patient stayed at the hospital during the initial admission. The length of stay was recorded as zero for patients who were not admitted to the hospital for at least 24 hours. Any GSI-related hospital stay after the first discharge was not included in this calculation. The total length of hospitalization was calculated based on the total number of days the patient was admitted to the hospital for a GSI of the hand-related treatment. The number of operations was defined as any surgery related to a GSI of the hand. The number of post-op visits was defined as the number of post-op visits at the hospital related to treatment of the GSI of the hand. The initial number of radiographic studies was defined as the number of radiographic studies needed during the first admission. The total number of radiographic studies was defined as the total number of radiographic studies performed for the treatment of a GSI of the hand. Radiographic studies included hand/wrist plain radiographs, computed tomography (CT) of the hand, magnetic resonance imaging (MRI) of the hand, and angiograms of the hand.

Statistical Analysis

Dichotomous and categorical data are presented as frequencies with percentages and continuous data as medians and interquartile ranges (IQRs) or means and standard deviations (SDs) based on normality. To assess the relationship between our explanatory variables and our continuous outcome variables (total length of hospitalization and number of operations), we used the Student’s T-test for dichotomous variables, Kruskal-Wallis test for categorical variables, and Spearman’s rank correlation for ordinal and continuous variables. The relationship between our explanatory variables and unplanned reoperation was assessed using the Fisher’s Exact for dichotomous variables and categorical variables and the Mann-Whitney-U test for ordinal and continuous variables. To mitigate confounding, variables with a P-value < .10 in bivariate analysis were entered into a multivariable model to identify factors independently associated with unplanned operation, total length of hospitalization, and number of operations. A P-value of < .05 was considered as statistically significant. All analyses were performed using STATA 16.0 (STATA Corp, College Station, Texas).

Results

Demographics

One hundred thirty-nine (93.9%) patients were male and the median age was 27.6 years (IQR: 22.0-41.8). Data regarding hand dominance and occupation were incomplete. However, based on the available data, 58.4% of the patients sustained an injury to the nondominant hand and the majority of patients were laborers (29.2%), unemployed (25.8%), or law enforcement officers/security (14.2%). Fifty-eight patients (39.2%) were referred from another hospital, of which 9 patients (15.5%) were transferred from another state and 3 patients (5.2%) were transferred from another country (Table 1). The last date of follow-up was based on the date of the last clinical visit to one of our hospitals, with the assumption that if further care was sought, this would occur at our institution if follow-up were ongoing. The median time from presentation to last date of follow-up was 1.0 year (IQR = 32 days to 4.7 years).

Table 1.

Study Population (n = 148).

Patient characteristics
Age (years), median (IQR)	27.6 (22-41.8)
Male sex, n (%)	139 (93.9)
Smoking, n (%)	58 (39.2)
Diabetes, n (%)	14 (9.5)
Occupation, n (%) a
Student	8 (6.7)
Skilled worker	10 (8.3)
Clerical work	17 (14.2)
Law enforcement and security	17 (14.2)
Laborer	35 (29.2)
Unemployed	31 (25.8)
Retired	2 (1.7)
Dominant hand affected, n (%) b	47 (41.6)
Referral status
Transfer from other hospital, n (%)	58 (39.2)
Transfer from another hospital in Massachusetts, n (%)	46 (79.3)
Transfer from out of state, n (%)	9 (15.5)
Transfer from another country, n (%)	3 (5.2)
Injury characteristics
Gun shot wound at other area, n (%)	55 (37.2)
Mechanism of injury, n (%)
Shot self	45 (30.4)
Shot by other	103 (69.6)
Wound characteristic, n (%)
Entry wound only	20 (13.5)
Entry wound and exit wound	120 (81.1)
Tangential wound	8 (5.4)
Retained bullet/fragment, n (%)	54 (36.5)
Associated fracture, n (%)	115 (77.7)
Associated nerve injury, n (%)	70 (47)
Associated arterial injury, n (%)	9 (6)
Associated tendon injury, n (%)	23 (15.5)
Associated compartment syndrome of hand, n (%)	3 (2)
Associated first web space injury, n (%)	17 (11.5)
Fracture severity
No fracture	33 (22.3)
Fracture without bone loss	56 (37.8)
Fracture with partial bone loss	30 (20.3)
Fracture with segmental bone loss	29 (19.6)

Note. IQR = interquartile range.

^aMissing data in 28 cases.

^bMissing data in 35 cases.

Injury Characteristics

Forty-five out of 148 patients (30.4%) had a self-inflicted injury. The type of gun that caused the injury was unknown in two-thirds of the medical chart, but 44 out of 48 (91.6%) of the GSI were caused by low-velocity pistol-type guns. Concomitant GSI outside the region of the hand were seen in 55 patients (37.2%) (Supplemental Figure 2). Most patients had a GSI with an exit wound, 20 patients (13.5%) had a GSI with only an entry wound, and 8 patients (5.4%) had a tangential wound. A retained bullet (fragment) was identified in 54 patients (36.5%). All 20 patients with only an entry wound had a retained bullet (fragment) and 30.9% of the patients (34 out 110) with an exit wound had a retained bullet fragment. One hundred and fifteen patients sustained a fracture in the hand (77.7%). The most common bone fractured was the metacarpal followed by the proximal phalanx (Supplemental Figure 3).

Nerve deficits occurred in 70 patients (47.3%) and ranged from neuropraxia to complete transection of the nerve. The digital nerve was involved in 58 patients (82.9%), ulnar nerve in 8 patients (11.4%), and median nerve in 4 patients (5.7%) (Table 2) (Supplemental Figure 4). Arterial injury occurred in 9 patients (6.1%) (Supplemental Figure 5).

Table 2.

Patients With Associated Nerve Deficit (n = 70).

Clinical and treatment characteristics of patients with a nerve deficit	n = 70
Clinical presentation	n (%)
Partial motor loss	1 (1.4)
Partial sensory/motor loss	4 (5.5)
Complete palsy	13 (18.6)
Nerve exploration	38 (54.3)
Intra-operative findings of nerve exploration (n=38)
Intact/contusion	16 (42.1)
Partial transection	5 (13.2)
Complete transection	17 (44.7)
Treatment for nerve injury
Direct repair	7 (10.0)
Repair with autograft	4 (5.7)
Repair with allograft	2 (2.9)
No repair	13 (18.6)
Neurolysis	12 (17.1)
Observation	32 (45.7)

Overall complication rate in our cohort was 16.9%, among which stiffness was most common (7.4%) (Supplemental Table 1). Among patients with fracture(s), there was a 7.8% rate of delayed or nonunion. The decision about delayed union and nonunion was based on surgeon description in medical record.

Treatment Characteristics

Ninety-five patients (64.2%) underwent surgical treatment for GSI of the hand, while the remainder were treated conservatively, most often with dressings and splinting. There were 37 patients (38.9%) who underwent an additional secondary operation for GSI of the hand, of which 27 were unplanned reoperations (Supplemental Table 2).

Bivariate analysis of unplanned reoperation can be found in Table 3. Multivariable logistic regression showed that only fracture severity was independently associated with unplanned reoperation (odds ratio: 1.71; P < .036) (Supplemental Table 3).

Table 3.

Factors Associated With Unplanned Reoperation.

	Unplanned reoperation		P-value
Patient and injury characteristics	Yes (n = 27)	No (n = 68)
Age (years), median (IQR)	34 (23-42)	32 (20-44)	.32
Male sex, n (%)	25 (93)	65 (96)	.62
Smoking, n (%)	11 (41)	29 (43)	1.00
Diabetes, n (%)	3 (11)	5 (7)	.68
Occupation, n (%) a			.71
Student	2 (9.1)	3 (5.5)
Skilled worker	3 (13.6)	10 (18.2)
Clerical work	2 (9.1)	5 (9.1)
Law enforcement and security	5 (22.7)	7 (12.7)
Laborer	6 (27.3)	14 (25.5)
Unemployed	3 (13.6)	15 (27.3)
Retired	1 (4.6)	1 (1.8)
Injury on dominant hand, n (%) b	8 (36.4)	22 (40.7)	.80
Mechanism of injury, n (%)			1.00
Shot self	9 (33.3)	23 (33.8)
Shot by other	18 (66.7)	45 (66.2)
Wound characteristics, n (%)			.84
Entry wound only	4 (14.8)	13 (19.1)
Entry and exit wound	23 (85.2)	54 (79.4)
Tangential wound	0 (0)	1 (1.5)
Retained bullet/fragment, n(%)	16 (59.3)	26 (38.2)	.07
Associated fracture, n(%)	25 (92.6)	62 (91.2)	1.00
Associated nerve injury, n(%)	17 (63.0)	40 (58.8)	.82
Associated arterial injury, n(%)	4 (14.8)	5 (7.4)	.27
Associated tendon injury, n(%)	7 (25.9)	16 (23.5)	.80
Associated compartment syndrome, n(%)	2 (7.4)	1 (1.5)	.19
Associated first web space injury, n(%)	4 (14.8)	9 (13.2)	1.00
Fracture severity			.04
1. No fracture	2 (7.4)	6 (8.8)
2. Fracture without bone loss	8 (29.6)	29 (42.7)
3. Fracture with partial bone loss	4 (14.8)	21 (30.9)
4. Fracture with segmental bone loss	13 (48.2)	12 (17.7)

Note. IQR = interquartile range.

^aData missing in 18 cases.

^bData missing in 19 cases.

Among 115 patients with associated hand fractures, 87 patients (75.7%) were treated surgically. Initial bone grafting was performed in 7 patients (6.1%) and delayed bone grafting in 12 patients (10.4%) (Supplemental Table 4).

Among the 70 patients with a nerve deficit during initial physical examination, 38 patients (54.3%) underwent surgical exploration. Seventeen of these surgical patients (44.7%) were found to have complete nerve transection, 5 patients (13.2%) were found to have partial nerve transection, and 16 patients (42.1%) were found to have an intact nerve (Table 2) (Supplemental Figure 4). Seven patients (10.0%) underwent direct nerve repair, 4 patients (5.7%) underwent nerve repair with autograft, 2 patients (2.9%) underwent nerve repair with allograft, 12 patients (17.1%) underwent neurolysis alone, and 13 patients (18.6%) did not undergo surgical intervention to the nerve. Thirty-two patients (45.7%) with nerve deficits did not undergo surgical exploration and were therefore categorized as “observation.”

Healthcare Utilization

Ninety-three patients (62.8%) with a GSI of the hand were admitted to the hospital, of which 2 were admitted for a second time. The median length of hospitalization was 1 day (IQR: 0-4; range: 0-30 days) for both the initial admission and the total length of hospitalization (Supplemental Table 5). Bivariate analysis regarding number of operations and total length of hospitalization can be found in Tables 4 and 5. Multivariable linear regression showed that only fracture severity is independently associated with a higher number of hand operations after a GSI of the hand (coefficient = 0.41; P = .01) (Supplemental Table 6). The total length of hospital stay was independently associated with gunshot injuries that had a retained bullet (fragment) (coefficient = 2.53; P = .01) and patients having gunshot injuries in other regions than the hand (coefficient = 4.34; P < .001) (Supplemental Table 7).

Table 4.

Patient and Injury Characteristics Associated With Increased Utilization of Hospital Resources; Bivariate Analysis for Number of Operations.

Characteristics (continuous variables)	Correlation coefficient		P—value
Age (years), median (IQR)	0.075		.37
Characteristics (dichotomous variables)	Median (IQR)		P—value
	Yes	No
Male gender	1 (0-2)	1 (0-1)	.84
Smoking	1 (0-1)	1 (0-2)	.81
Diabetes	1 (0-1)	1 (0-2)	.58
Injury on dominant hand a	1 (0-2)	1 (0-2)	.62
Gun shot wound at other area	1 (0-2)	1 (0-1)	.96
Retained bullet/fragment	1 (1-2)	1 (0-1)	<.01
Associated fracture	1 (1-2)	0 (0-0)	<.001
Associated tendon injury	1 (1-2)	1 (0-1)	<.01
Associated artery injury	2 (1-2)	1 (0-1)	<.01
Associated compartment syndrome	2 (1-4)	1 (0-1)	.08
Associated nerve injury	1 (1-2)	0 (0-1)	<.001
Associated first web space injury	1 (1-2)	1 (0-1)	.32
Characteristics (categorical variables)	Median (IQR)		P-value
Occupation b			.36
Student	1 (0-2)
Skilled worker	1 (0-2)
Clerical work	1 (1-2)
Law enforcement and security	1 (0-2)
Laborer	1 (0-1)
Unemployed	1 (0-1)
Retired	2.5 (2-3)
Mechanism of injury			.37
Shot self	1 (0-2)
Shot by other	1 (0-1)
Wound characteristics			<.01
Entry wound only	1 (1-2)
Entry and exit wound	1 (0-2)
Tangential wound	0 (0-0)
Characteristics (ordinal variables)	Median (IQR)		P-value
Fracture severity			<.001
1. No fracture	0 (0-0)
2. Fracture without bone loss	1 (0-1)
3. Fracture with partial bone loss	1 (1-1)
4. Fracture with segmental bone loss	1 (1-3)

Note. IQR = interquartile range. Values in bold indicate statistical significance at p<0.05.

^aMissing data in 35 cases.

^bMissing data in 28 cases.

Table 5.

Patient and Injury Characteristics Associated With Increased Utilization of Hospital Resources; Bivariate Analysis for Total Length of Hospitalization.

Characteristics (continuous variables)	Correlation coefficient		P-value
Age (years), median (IQR)	–0.13		.21
Characteristics (dichotomous variables)	Median (IQR)		P-value
	Yes	No
Male gender	1 (0-4)	3 (0-7)	.26
Smoking	1 (0-4)	1 (0-4)	.30
Diabetes	0.5 (0-4)	1 (0-4)	.71
Injury on dominant hand a	1 (0-4)	1 (0-3)	.93
Gun shot wound at other area	4 (1-7)	0 (0-2)	<.001
Retained bullet/fragment	1.5 (0-7)	1 (0-3)	<.01
Associated fracture	1 (0-4)	0 (0-3)	<.05
Associated tendon injury	1 (1-4)	1 (0-4)	.29
Associated artery injury	4 (3-4)	1 (0-4)	<.01
Associated compartment syndrome	2 (2-4)	1 (0-4)	.28
Associated nerve injury	1 (0-4)	1 (0-3)	.11
Associated 1st web space injury	1 (0-4)	1 (0-4)	.95
Characteristics (categorical variables)	Median (IQR)		P-value
Occupation b			.74
Student	2 (0-6)
Skilled worker	1.5 (1-7)
Clerical work	1 (0-4)
Law enforcement and security	1 (0-2)
Laborer	1 (0-3)
Unemployed	0 (0-4)
Retired	2.5 (2-3)
Mechanism of injury			.30
Shot self	1 (0-2)
Shot by other	1 (0-5)
Wound characteristics			.13
Entry wound only	1 (0-4)
Entry and exit wound	1 (0-4)
Tangential wound	0 (0-0.5)
Characteristics (ordinal variables)	Median (IQR)		P-value
Fracture severity			.02
1. No fracture	0 (0-3)
2. Fracture without bone loss	1 (0-4)
3. Fracture with partial bone loss	1 (0-4)
4. Fracture with segmental bone loss	1 (1-4)

Note. IQR = interquartile range. Values in bold indicate statistical significance at p<0.05.

^aMissing data in 35 cases.

^bMissing data in 28 cases.

Discussion

In this retrospective study, we aimed to evaluate the patient, injury, and treatment characteristics of patients who had a GSI of the hand and to assess the utilization of healthcare resources in patients with a GSI of the hand. Of the patients that presented at one of our centers with a GSI of the hand: 64% had surgical treatment of which 28% had unplanned reoperation, 47% had a nerve deficit, 37% had a retained bullet fragment, and 30% were reported to be self-inflicted injuries. GSI of the hand with more complex fractures were associated with unplanned reoperation and higher number of operations, while a retained bullet (fragment) and GSI in other location were associated with increased length of hospitalization.

The results of this study need to be viewed in the light of its limitations. First, the use of ICD-9 and 10 codes to identify cases may lead to selection bias due to incorrect coding. We aimed to reduce this bias by using multiple ICD codes that could be used for GSI’s of the hand and manually reviewing the medical chart of patients with one of these codes (Supplemental Material 1). Second, this is retrospective study over 17 years which makes it one of the largest cohort studies for hand GSIs. However, medical chart data can be incomplete or inaccurate, causing nondifferential misclassification bias. Third, it is possible that patients had their follow-up at another hospital after they were initially treated at one of our centers. Low compliance rates in patients with GSIs to the hand have been reported with up to 39% lost to follow-up after initial hospital discharge. ⁸ It is therefore possible that the median number of operations and length of hospitalization is underestimated. Fourth, results of this study may not be generalizable to all centers, as some patients were likely transferred for other life-threatening injuries or because of the need for microsurgical expertise.

According to the Centers for Disease Control and Prevention, firearm-related violence accounted for 39,773 overall deaths and 14,542 assault-related deaths in the United States in 2017, affecting males with an age-adjusted death rate 6.1 times greater than females. ⁹ Analysis of individual U.S. states reveals a wide range of firearm mortality rates across the country, including 3.5 per 100,000 deaths in Massachusetts and up to 22.9 per 100,000 in Mississippi. ¹⁰ GSIs contribute to a large number of orthopedic and soft-tissue-related injuries and are a leading cause of extremity fractures and disability in the United States. ³ With an estimated annual cost of $2.1 billion, nonfatal GSIs also continue to place a significant burden on the healthcare system. ⁷ The incidence of nerve deficit after a GSI in our cohort is relatively high (47%) compared to other studies reporting an incidence ranging from 8% to 43%.^6,11,12 This difference may be declared by the different definitions these studies use for nerve injury and the fact that some of these studies did not focus on solely GSI of the hand. In our study, the digital nerve was the most commonly injured nerve followed by ulnar nerve and median nerve, which is similar to a prior report. ⁶

It is commonly posited that most nerve deficits after a GSI are the result of blast effect and should therefore recover spontaneously. In addition, a prior study reported that 70% of the nerve injuries after a GSI of the upper extremity recover spontaneously within 9 months.^13,14 In our cohort, 22 out of 148 patients (15% of total patients) were found to have either complete or partial nerve transection during surgical exploration. A prior study examining predictors of nerve injury in patients who had a GSI of the upper extremity reported 40 out of 123 patients (33%) had an exploration of the nerve. Their results showed that a complete palsy was a predictor for nerve transection.

Multivariable analysis demonstrated that a retained bullet (fragment) in the hand after a GSI was associated with longer hospital stay and showed a trend toward more unplanned reoperation as some of these patients underwent additional surgery for removal of a bullet fragment after initial treatment. A retained bullet (fragment) in the hand is a relative indication for removal as they might impede joint motion and are often easily removed when located in the subcutaneous tissue.^3,15 There are no accepted criteria for bullet fragment removal, but some include skin compromise, symptoms related to the retained fragment, and patient preference. All patients who did not have an exit wound and 30% of the patients with an exit wound had a retained bullet (fragment). A retained bullet (fragment) usually occurs when the bullet hits cortical bone. If this occurs, the bullet’s remaining kinetic energy is transferred to the bone and the surrounding tissue, which may result in a higher degree of soft tissue injury. However, if the bullet is projected by a high-velocity weapon and has enough kinetic energy it will even traverse cortical bone, which can lead to even more severe damage.^13,16 This may indicate why the type of wound did not show any association with any of our outcomes, whereas fracture severity is associated with unplanned reoperation and a higher number of surgical procedures.

Another factor that was independently associated with longer hospital stay is a concomitant GSI outside of the region of the hand. These patients likely need a longer hospital stay due to nonhand injuries that were treated at the same time as their hand injury.

Multivariable analysis demonstrated that fracture severity was associated with a higher number of operations. While patients without fracture or a simple fracture can be treated nonoperatively, fractures with bone loss may require surgical treatment with or without bone graft.^13,17 There is still a controversy about the use of bone graft initially or in a staged procedure. ³ Some studies suggest that initial bone grafting may increase the incidence of infection because these type of injuries are often accompanied by soft tissue injury.^18,19 In our study, 19 patients (16.5%) had a bone graft, of which 12 were staged. Due to the low number of patients with bone loss, we could not estimate if initial or delayed bone grafting is associated with infection.

In this study, 30% of the patients reported they shot themselves accidentally. A prior study reported that 81% of the GSIs to the hand were self-inflicted, and in more than 60% of the time caused by cleaning or loading the gun. ⁶ Other self-inflicted GSIs may not be accidental. We observed 32% of the self-inflicted trauma occurred in patients working in the law enforcement and security sector. Although these patients are more trained in the use of a gun, they are also more exposed to the risks of owning and using a gun.

Conclusions

Approximately 6% of all patients that presented with a GSI at one of our hospitals sustained GSIs of the hand. Some degree of nerve deficit is common in GSIs of the hand with 47% having sensory and/or motor symptoms and up to 15% having a nerve transection. Fracture severity and a retained bullet fragment were associated with unplanned reoperation, a higher number of operations, and longer stay at the hospital. These findings are helpful for patient counseling and understanding the impact of GSIs of the hand on the health care system.