The Burden of Liver Trauma in an Urban Trauma Centre in Johannesburg, South Africa

1. Introduction

Liver trauma is very common following blunt and penetrating trauma to the lower chest and upper abdomen [1,2,3]. The liver is the most commonly injured organ in blunt abdominal trauma and the second most commonly injured organ in penetrating abdominal trauma [3,4,5,6]. The severity of penetrating liver injury depends on the trajectory of the missile or implement. Injuries can range from simple lacerations of the parenchyma to major vascular injuries [3].

Operative intervention for the management of liver injuries is needed in approximately 14% of patients, according to international literature. These are generally patients who initially present with hemodynamic instability, who fail non-operative management or require surgery for other intra-abdominal injuries [2,3,7,8,9].

Associated injuries to other organs occur in approximately 80% of patients with hepatic injury. The type of associated injuries is generally dependent on the mechanism of trauma. Patients with blunt trauma can have involvement of multiple organ systems, whereas in penetrating trauma, it is mostly surrounding structures that are injured [4,5,6].

The strategy used to manage liver injury depends on the patient's haemodynamic status, grade of liver injury, presence of other injuries, and medical comorbidities.

The haemodynamically unstable blunt trauma patient with a positive Focused Assessment with Sonography for Trauma (FAST) scan requires an emergent exploratory laparotomy to determine the source of intraperitoneal bleeding [6,9,10,11,12]. Patients with blunt liver injury who are haemodynamically stable and do not have other indications for exploratory laparotomy exploration can be observed. Haemodynamically stable patients with right-sided penetrating thoracoabdominal injuries that lacerate the liver can also be observed, provided that there are no associated intra-abdominal injuries. If there is any suspicion of an associated intra-abdominal injury in a penetrating trauma patient, based on the trajectory of the missile, clinical findings of peritonism, or scan findings, emergent laparotomy is mandated. Computed tomography (CT) is the preferred imaging modality to assess the severity of liver trauma and the presence or absence of other intra-abdominal injuries [6,7,9,10,11,13].

Complications related to liver trauma are most frequently biliary complications – bilomas, bile leaks, biliary peritonitis, and bile sepsis, which are estimated to occur in approximately 3.2% of all blunt liver trauma patients. In penetrating liver trauma, the complication rate has been reported to be 50 – 52% [6,7,8,14].

This study aimed to examine the burden of liver trauma in a busy urban trauma centre in South Africa. What makes this patient population unique is the significant number of patients who sustained liver injury due to penetrating injuries and the large number of patients who underwent surgical interventions. This is vastly different from the bulk of international data, which focuses on non-operatively managed blunt liver trauma.

2. Patients and Methods

Data were gathered retrospectively from all priority one patients presenting with liver trauma over a seven-year period from 1 January 2017 to 31 December 2023 at the Trauma Unit of Charlotte Maxeke Johannesburg Academic Hospital in Johannesburg, South Africa. A priority one (P1) patient was defined as a critically ill or injured patient requiring immediate attention, or an unstable patient with a potentially life-threatening injury or illness.

Data were collected regarding demographics, mechanism of injury, grade of liver injury, associated injuries to other body regions, associated intra-abdominal injuries, Injury Severity Score (ISS), New Injury Severity Score (NISS), need for emergency surgery, and mortality.

Ethical approval was obtained from the Human Research Ethics Committee of the University of the Witwatersrand.

All statistical analyses were conducted using Stata (V.18, Stata Corp, College Station, Texas, USA), R (version 4.3.2), and RStudio (2023.9.0.463) software. Continuous variables were expressed as means with standard deviations for normally distributed data and as median with interquartile range (IQR) for non-normally distributed data. Categorical variables were expressed as frequencies and percentages. Chi-square and Wilcoxon rank-sum tests were used to evaluate differences between categorical variables and differences in medians for continuous variables with p-values. The Wald test and Cox proportional hazard models were used to compute the statistical significance (p-value) of the differences between Kaplan-Meier survival curves and hazard ratios. Statistical significance was set at P < 0.05.

The severity of liver injury is reported using the organ injury scale (OIS) described by the American Association for the Surgery of Trauma (AAST) and were graded I – V based on CT scan or operative findings.[15]

3. Results

3.1. Demographics

Data were collected from 512 patients with complete medical records. The median age of the patients was 31 years (26 – 36). The majority of the patients (466, 91%) were male, and only 46 (9%) were female. Most patients were of African ethnicity.

Table 1. Demographic data of liver trauma patients.

Category	Characteristics	Number (%) or Median (IQR: Q1-Q3) n = 512
Age		31 (26-36)
Gender
	Male	466 (91)
	Female	45 (9)
Ethnicity
	African	484 (95)
	Coloured	4 (1)
	Indian	4 (1)
	Caucasian	3 (0.59)
	Asian	2 (0.39)
	Hispanic	1 (0.20)

Table 1. Demographic data of liver trauma patients.

Category	Characteristics	Number (%) or Median (IQR: Q1-Q3) n = 512
Age		31 (26-36)
Gender
	Male	466 (91)
	Female	45 (9)
Ethnicity
	African	484 (95)
	Coloured	4 (1)
	Indian	4 (1)
	Caucasian	3 (0.59)
	Asian	2 (0.39)
	Hispanic	1 (0.20)

3.2. Clinical Characteristics

Penetrating trauma to the liver was much more common than blunt trauma (73% vs. 27 %). One patient experienced a combination of blunt and penetrating traumas. Of the 373 patients who sustained penetrating trauma, 212 (57%) sustained gunshot wounds, 160 (43%) were stabbed, and a sharp object impaled one.

In the blunt trauma cohort, the most common mechanism was road traffic collision, which accounted for 113 patients (82%). These were further subdivided into motor vehicle or motorcycle collisions, with 57 patients (41%) sustaining liver injuries secondary to this mechanism, and pedestrian-vehicle collisions accounting for liver injuries in 56 patients (41%). With respect to other blunt causes, 12 patients (9%) sustained a liver injury due to a fall from a height, blunt assault was the mechanism of injury in eight patients (6%), and five patients (4%) sustained liver trauma secondary to a heavy object falling on them.

Regarding the grade of liver injury, 65 patients (13%) sustained grade I liver injury, 171 patients (33%) sustained grade II liver injury, 187 patients (37%) sustained grade III liver injury, 79 patients (15%) sustained grade IV liver injury, and 10 patients (2%) sustained grade V liver injury.

The average ISS was 14 and the NISS was 19. The overall mortality rate was 5 %; blunt liver trauma patients had a 7% mortality rate compared to 5% in the penetrating trauma group (p=0.796). The mortality rate was 10% in the group that required emergency surgery compared with 1% in the group of patients managed non-operatively (p < 0.001).

Emergency laparotomy was performed in 261 (51 %) patients. Among penetrating trauma patients, 232 (62%) required surgery, while only 28 (20%) patients with blunt liver trauma required surgery (p<0.0001). A single patient with a combined mechanism of injury also underwent laparotomy.

Table 2. Clinical data of liver trauma patients.

Category	Characteristics	Number (%) or Median (IQR: Q1-Q3) n = 512
Blunt or penetrating mechanism
	Penetrating	373 (73)
	Blunt	138 (27)
	Combination	1 (0.20)
Mechanism of injury
	GSW	212 (41)
	Stab	160 (31)
	MVC/MBC	57 (11)
	PVC	56 (11)
	Fall from height	12 (2)
	Assault	8 (2)
	Heavy object falling on patient	5 (1)
	Impaled	1 (0.2)
Grade of liver injury
	I	65 (13)
	II	171 (33)
	III	187 (37)
	IV	79 (15)
	V	10 (2)
ISS		14 (9-22)
NISS		19 (11-29)
Survival
	Alive	484 (95)
	Dead	28 (5)
Surgery
	Yes	261 (51)
	No	251 (49)

MVC – Motor Vehicle Collision, MBC – Motorbike Collision, PVC- Pedestrian Vehicle Collision, GSW – Gunshot Wound.

Table 2. Clinical data of liver trauma patients.

Category	Characteristics	Number (%) or Median (IQR: Q1-Q3) n = 512
Blunt or penetrating mechanism
	Penetrating	373 (73)
	Blunt	138 (27)
	Combination	1 (0.20)
Mechanism of injury
	GSW	212 (41)
	Stab	160 (31)
	MVC/MBC	57 (11)
	PVC	56 (11)
	Fall from height	12 (2)
	Assault	8 (2)
	Heavy object falling on patient	5 (1)
	Impaled	1 (0.2)
Grade of liver injury
	I	65 (13)
	II	171 (33)
	III	187 (37)
	IV	79 (15)
	V	10 (2)
ISS		14 (9-22)
NISS		19 (11-29)
Survival
	Alive	484 (95)
	Dead	28 (5)
Surgery
	Yes	261 (51)
	No	251 (49)

MVC – Motor Vehicle Collision, MBC – Motorbike Collision, PVC- Pedestrian Vehicle Collision, GSW – Gunshot Wound.

3.3. Associated Injuries

Of all patients with liver injuries, the majority (306 patients, 60%) had associated injuries.

Associated injuries to other body regions occurred in 80% of blunt trauma patients and 53% of penetrating trauma patients (p<0.0001). A single patient with a combination of blunt and penetrating trauma had an associated traumatic brain injury.

The most common body area with associated injuries was the chest; overall, 30% of the patients sustained associated chest injuries. This was true for both blunt and penetrating trauma subgroups. Associated soft tissue injuries requiring surgical intervention occurred in 77 patients (15%); the majority of these were patients who experienced penetrating trauma. Twenty-five patients (4.8%) had associated traumatic brain injuries, the vast majority of which were in the blunt trauma group. Twenty-two patients (4.3%) had associated long bone fractures, 17 (3.3%) had associated spinal fractures, and 9 (1.8%) had associated pelvic fractures. Associated neck injuries were uncommon and were only observed in three patients (1%).

Associated intra-abdominal injuries occurred in 37% of patients with blunt mechanisms of injury and 64% of patients with penetrating trauma (p<0.0001). A single patient with a combination of blunt and penetrating trauma had associated injury to the large bowel.

A total of 289 associated abdominal injuries were observed. In total, 114 (22%) patients had injuries to the abdominal wall or mesentery, requiring intervention. There were 51 patients (10%) with associated renal injuries, 45 (8.8%) had associated large bowel injuries, 41 (8%) had associated splenic injuries, and 28 (5.5%) had associated small bowel injuries. Intra-abdominal vascular injuries occurred in 10 (2 %) patients.

Associated injuries to the spleen were the most common in the blunt trauma subgroup, while injuries to the mesentery or abdominal wall requiring repair were the most common in the penetrating trauma subgroup.

3.4. Complications

There were 34 complications in 33 patients; thus, the complication rate was 6.6% (one patient had two complications). Among the total number of complications, localised bile leak was the most common complication in 14 patients (2.7%), followed by abscess formation in seven patients (1.4%). Vascular complications occurred in 6 patients (1.2%); biliopleural fistulae occurred in 3 patients (0.6%); these were all patients who sustained penetrating trauma to the right upper quadrant. Three patients (0.6%) developed necrosis of the injured portion of the liver, one patient (0.2%) developed a very large subcapsular haematoma, and one patient (0.2%) developed biliary peritonitis.

Complications occurred more commonly in patients who required surgical management of their liver injuries (11.1% vs. 2%, p < 0.0001). Complications were more common in patients who sustained penetrating trauma (7.8%) than in those who sustained blunt trauma (3.6%) but not statistically significant (p = 0.09).

Table 3. Associated injuries and complications.

Associated injuries of other body regions
Body region	Number (%) n = 512	Blunt trauma n = 138	Penetrating trauma n = 373	Combination n = 1
Chest injury	153 (30)	44 (32)	109 (29)	0
Soft tissue injury requiring surgical intervention	77 (15)	4 (2.9)	73 (20)	0
Traumatic brain injury	25 (4.8)	23 (17)	1 (0.3)	1 (100)
Long bone injury	22 (4.3)	19 (14)	3 (0.8)	0
Spine injury	17 (3.3)	7 (5)	10 (2.7)	0
Pelvic injury	9 (1.8)	9 (6.5)	0	0
Neck injury (excluding C-spine)	3 (0.6)	3 (2.2)	0	0
Associated intra-abdominal injuries
Organ injured	Number (%) n = 512	Blunt trauma (%) n = 138	Penetrating trauma (%) n = 373	Combination (%) n = 1
Abdominal wall/mesentery requiring intervention	114 (22)	5 (3.6)	109 (29)	0
Kidney	51 (10)	17 (12)	34 (9.1)	0
Large bowel	45 (8.8)	3 (2.2)	42 (11)	1 (100)
Spleen	41 (8)	24 (17)	17 (4.6)	0
Small bowel	28 (5.5)	3 (2.2)	25 (6.7)	0
Intra-abdominal vascular	10 (2)	0	10 (2.9)	0
Complications related to liver trauma
Complication	Total (%) n = 512	Blunt trauma (%) n = 138	Penetrating trauma (%) n = 373	Combination (%) n = 1
Total	34 (6.6)	5 (3.6)	29 (7.8)	0
Localised bile leak	14 (2.7)	0	14 (3.8)	0
Abscess	7 (1.4)	0	7 (1.9)	0
Vascular	6 (1.2)	3 (2.2)	3 (0.8)	0
Necrosis	3 (0.6)	1 (0.7)	2 (0.5)	0
Biliopleural fistula	3 (0.6)	0	3 (0.8)	0
Biliary peritonitis	1 (0.2)	1 (0.7)	0	0

Table 3. Associated injuries and complications.

Associated injuries of other body regions
Body region	Number (%) n = 512	Blunt trauma n = 138	Penetrating trauma n = 373	Combination n = 1
Chest injury	153 (30)	44 (32)	109 (29)	0
Soft tissue injury requiring surgical intervention	77 (15)	4 (2.9)	73 (20)	0
Traumatic brain injury	25 (4.8)	23 (17)	1 (0.3)	1 (100)
Long bone injury	22 (4.3)	19 (14)	3 (0.8)	0
Spine injury	17 (3.3)	7 (5)	10 (2.7)	0
Pelvic injury	9 (1.8)	9 (6.5)	0	0
Neck injury (excluding C-spine)	3 (0.6)	3 (2.2)	0	0
Associated intra-abdominal injuries
Organ injured	Number (%) n = 512	Blunt trauma (%) n = 138	Penetrating trauma (%) n = 373	Combination (%) n = 1
Abdominal wall/mesentery requiring intervention	114 (22)	5 (3.6)	109 (29)	0
Kidney	51 (10)	17 (12)	34 (9.1)	0
Large bowel	45 (8.8)	3 (2.2)	42 (11)	1 (100)
Spleen	41 (8)	24 (17)	17 (4.6)	0
Small bowel	28 (5.5)	3 (2.2)	25 (6.7)	0
Intra-abdominal vascular	10 (2)	0	10 (2.9)	0
Complications related to liver trauma
Complication	Total (%) n = 512	Blunt trauma (%) n = 138	Penetrating trauma (%) n = 373	Combination (%) n = 1
Total	34 (6.6)	5 (3.6)	29 (7.8)	0
Localised bile leak	14 (2.7)	0	14 (3.8)	0
Abscess	7 (1.4)	0	7 (1.9)	0
Vascular	6 (1.2)	3 (2.2)	3 (0.8)	0
Necrosis	3 (0.6)	1 (0.7)	2 (0.5)	0
Biliopleural fistula	3 (0.6)	0	3 (0.8)	0
Biliary peritonitis	1 (0.2)	1 (0.7)	0	0

4. Discussion

The protocol for management of liver trauma patients in this unit was initial assessment and resuscitation according to ATLS® principles.[16] Patients with blunt trauma who had free fluid on FAST examination and were haemodynamically unstable or patients with peritonism underwent exploratory laparotomy without the routine performance of a CT scan. Stable patients and those who responded to initial fluid resuscitation underwent a CT scan where the grade of the liver injury and the presence of other intra-abdominal injuries were determined. Isolated solid organ injuries in haemodynamically stable patients are routinely managed nonoperatively. At our institution, emergency access to interventional radiology was not available over weekends and after hours and rarely during working hours; therefore, all unstable patients underwent surgery.

Patients with penetrating abdominal trauma with evidence of peritoneal penetration, haemodynamic instability, or peritonitis underwent exploratory laparotomy without the routine performance of a CT scan. The only exception to this rule would be a haemodynamically stable patient without peritonism, who has an isolated penetrating wound in the right upper quadrant. If the CT scan showed no signs of injury to other intra-abdominal organs, these patients were managed nonoperatively. All patients with high-grade liver injuries (grades IV and V) were observed in the ICU setting.

Most patients who sustained liver trauma were of African ethnicity, which correlates with the demographics of the population living in the catchment area of Charlotte Maxeke Johannesburg Academic Hospital. Liver trauma was also mainly seen in male patients which mirrors the overall trend of trauma being more common in males [17,18,19,20,21]. The median age of patients who sustained liver trauma in this study was 31 years. This again mirrors global trends, where a disproportionate number of traumatic injuries occur in young males, with trauma being the leading cause of death in patients younger than 35 years [21,22].

Blunt trauma is the most common cause of liver trauma globally, with literature from Europe, the United Kingdom, the Far East, India, and Scandinavia showing a preponderance of blunt hepatic trauma [21,22,23,24,25,26,27,28,29,30]. Guidelines from the Eastern Association for the Surgery of Trauma and the Western Trauma Association in the United States focus solely on blunt hepatic trauma, with no similar guidelines on penetrating liver trauma published by these societies [6,9]. Our study showed a significant preponderance of penetrating liver trauma, which is very different from the global trends. The only other region in the world that seems to have a preponderance of penetrating rather than blunt liver trauma is South America [31,32].

South Africa has long been known as one of the most violent countries in the world, with a very high incidence of interpersonal and gun violence. Efforts have been made to make access to firearms stricter by introducing the Firearms Control Act (FCA) (Act 69, 2000). This was initially very successful in reducing the rate of firearm-associated homicides; however, since 2011, the firearm-related mortality rate has been steadily increasing again which is due to a combination of relaxation of the strict licencing conditions of the FCA and police corruption that allowed increased availability of unlicensed firearms [33,34].

In the penetrating trauma subgroup, gunshot wounds were more prevalent than were stab wounds. Most blunt trauma cases occur secondary to road traffic collisions. When looking at the incidence of liver trauma secondary to interpersonal violence, by combining the injuries secondary to stabbing, gunshots and assault, 74% of patients sustained liver trauma secondary to interpersonal violence. The predominance of Grade I-III injuries (83%) mirrors the international findings [10,12,23,24,32].

The fact that the average NISS was higher than the median ISS (19 versus 14) may, in part, be because 56% of the patients sustained associated abdominal injuries and there was a high incidence of penetrating trauma in our study (73%).

The all-cause mortality rate was 5% in the study group. This mortality rate is on par with, and in some cases, lower than, what is reported in a number of local and international studies and reflects well on the appropriateness of our institution’s approach to liver trauma [35,36,37,38,39]. Although there was a slightly higher mortality rate in the penetrating trauma group than in the blunt trauma group (7% vs. 5%), the difference was not statistically significant (p=0.796). The mortality rate was 10% in the group that required emergency surgery compared with 1% in the group of patients managed non-operatively (p < 0.001). There could be many factors that contribute to this, including the fact that these patients may have been more physiologically unwell or had more severe liver injuries. Further subgroup analyses regarding factors contributing to mortality need to be conducted to draw meaningful conclusions.

Modern management of liver trauma is increasingly moving to non-operative management in haemodynamically stable patients, and good guidelines exist regarding selective non-operative management of blunt liver trauma [6,7,9,37]. There is also a significant shift happening towards selective non-operative management of penetrating liver trauma provided that the patient is stable, the patient is not peritonitic and high-quality cross-sectional imaging excludes hollow viscus injury [14,27,38,40,41,42]. Some guidelines also recommend that selective nonoperative management should only be considered in low-energy penetrating trauma (stab wounds and gunshots due to handguns) [7].

The proportion of patients with liver injuries who require emergency laparotomy is not always easy to determine because many studies do not differentiate between blunt and penetrating injuries. However, this may be because penetrating trauma is rare in certain parts of the world. In mixed cohorts, surgery is required in 20 – 30% of patients who sustain liver trauma. When comparing the need for surgery in blunt versus penetrating liver trauma, 23 – 32% of patients with blunt liver trauma and 74 – 83% of patients who sustained penetrating liver trauma required surgery [35,36,37,38]. In our study, 51% of patients required laparotomies. More patients with penetrating liver trauma required surgery (62%), which was a statistically significant difference from those who sustained blunt liver trauma. This compares favourably with the rate of emergency surgery described in other sources, as stated above. The higher rate of surgery for penetrating trauma is possibly linked to the fact that associated intra-abdominal injury occurred in 64% of the patients who sustained penetrating liver trauma.

Associated injuries to other organ systems are frequently seen in many patients who have sustained blunt liver injuries, as these are often polytraumatised [23,25,35]. This finding was mirrored in our data, where 80% of blunt liver trauma patients sustained a concomitant injury to at least one other organ system compared to 53% of penetrating liver trauma patients who sustained a concomitant injury to another organ system (p < 0.0001). The high number of associated injuries in the penetrating trauma subgroup was mainly due to chest injuries resulting from either multiple penetrating injuries or a thoracoabdominal penetrating injury, as well as soft tissue injuries requiring surgical intervention. Injuries to organ systems other than these were rare in the penetrating trauma subgroup.

Injuries to other abdominal solid organs were the most common associated intra-abdominal injuries in the blunt trauma subgroup, with very few injuries to a hollow viscus. Although associated abdominal wall or mesenteric injuries requiring intervention were the most common injuries in the penetrating trauma group, hollow visceral injuries were also quite common in this group. All associated intra-abdominal vascular injuries occurred in the penetrating trauma group.

Biliary complications were the most common, occurring in 3.5% of patients in our study, with an overall complication rate of 6.6%. The most common complications related to liver trauma are biliary complications including bilomas, bile leaks, biliary peritonitis, and bile sepsis. These are estimated to occur in approximately 3.2 - 7% of all patients with blunt liver trauma. In penetrating liver trauma, the complication rate has been reported to be 50 – 52%. Intrahepatic abscesses can occur after liver trauma, mostly in high-grade injuries, at a rate of 0.6 – 7%. Our results were comparable to those reported in the literature. Vascular complications in the form of pseudoaneurysms or arteriovenous fistulae, as well as large subcapsular haematomas that may lead to secondary haemorrhage, are less common at a rate of 1.7 – 5.9%. Hepatic necrosis is a rare complication, as noted in only three in our cohort. In general, some studies estimate that complications occur in 12 – 14% of patients after liver trauma [6,7,8,14,36,43].

Complications occurred significantly more frequently in the patients who underwent surgical management of liver trauma. This may be because these patients were more likely to have higher-grade liver injuries, but further analysis of the data needs to be performed before any conclusions can be made. Complications occurred more than twice as commonly in patients who sustained penetrating trauma, but the difference was not statistically significant.

A potential explanation for why all cases of localised bile leaks were reported in the group of patients who suffered penetrating trauma may be that these patients underwent surgery where drains were left around the liver injury. This means that even a small bile leak would have been identified in contrast with non-operatively managed blunt trauma patients, where small bile leaks that resolve spontaneously would not be diagnosed. This needs to be further investigated to reach any conclusions.

5. Conclusions

In conclusion, this study provides valuable insights into the management and outcomes of liver trauma in a low-middle-income country. The findings revealed a unique preponderance of penetrating liver trauma, reflecting the high rates of interpersonal violence in South Africa. Despite this challenging context, the mortality rate of 5% is comparable to or lower than the international standards, indicating the effectiveness of the institution's approach. This study highlights the successful implementation of non-operative management in suitable cases, while also demonstrating the necessity for surgical intervention in a significant proportion of patients, particularly those with penetrating injuries. The complication rates observed are favourable compared to those reported in the international literature, with biliary complications being the most common. These results underscore the importance of tailored management strategies based on specific injury patterns and patient demographics across different regions. Further research is warranted to explore the factors contributing to mortality and complications, which could lead to improvements in liver trauma management protocols.

Limitations of the Study

This was a retrospective single-centre study, which was open to bias. There was no long-term follow-up of the cases, and some late complications may have been missed. The high prevalence of penetrating mechanisms may have introduced bias into the data. There is a need for a multicentre prospective study on liver trauma in a low-middle-income environment.