Symptomatic pseudoaneurysms following laparoscopic cholecystectomy: Focus on an unusual and dangerous complication.

BACKGROUND: Laparoscopic cholecystectomy (LC) has been associated with an increase in the incidence of biliary and vascular injuries. Pseudoaneurysms (PAs) following LC are rare life-threatening events with limited available experience regarding diagnosis and treatment.
MATERIALS AND METHODS: An extensive review of literature during a 26-year period (1994-2020) using MEDLINE® database and Google Scholar® academic search engine revealed 134 patients with at least one symptomatic PA following LC.
RESULTS: Nearly. 81% of patients with PAs become symptomatic during the first 8 weeks following LC. The most common symptoms were gastrointestinal bleeding (74%) and abdominal pain (61%). In 28% of cases, there was a concomitant bile duct injury or leak from the cystic duct stump, whereas in about one-third of cases, PAs presented following an uneventful LC. The most common involved arteries were the right hepatic artery (70%), the cystic artery (19%) or both of them (3%). Trans-arterial embolisation was the favoured first-line treatment with a success rate of 83%. During a median follow-up of 9 months, the mortality rate was 7%.
CONCLUSION: Clinicians should be aware of the PA occurrence following LC. Prompt diagnosis and treatment are essential.

INTRODUCTION

Prof. Dr Med Erich Mühe performed the first laparoscopic cholecystectomy (LC) in Germany in 1985.[1] Nowadays, LC is the gold standard treatment for symptomatic cholelithiasis and one of the most commonly performed abdominal operations worldwide. However, LC is associated with higher risk of biliary and vascular injuries compared with the conventional open procedure. Vascular injuries are less common than biliary injuries, but they carry a higher mortality rate.[2] Vascular injuries may lead to perioperative bleeding, hepatic artery infarction or pseudoaneurysm (PA) formation.[3] Although PAs are considered a rare complication following LC, their actual incidence may be much higher.[4] Vascular injuries that lead to PA formation are often not recognised at the time of surgery. Furthermore, PAs are characterised by an insidious and unpredictable clinical course, and they can cause severe life-threatening haemorrhage weeks to months after the LC.[5] This article constitutes an extensive review of the literature and summarises the published experience regarding PAs following LC. Two previous reviews of the literature have been published in the past which provided valuable information on this subject.[67]

MATERIALS AND METHODS

We conducted a literature search using the MEDLINE® database (National Library of Medicine, Bethesda, MD, USA) and Google Scholar® academic search engine (last search: 1 June, 2020) for published studies describing PA development following LC. We used the terms ‘pseudoaneurysms AND laparoscopic cholecystectomy’ and ‘false aneurysms AND laparoscopic cholecystectomy’. Inclusion criteria were (1) original reports with at least one patient, (2) written in English language and (3) describing the occurrence of at least one symptomatic PA following LC. We included patients who underwent LC regardless of whether it was converted to open or not. In addition, (4) the formation of PAs should be attributed to the initial surgical operation and not to possible post-operative interventions (e.g., endoscopic retrograde cholangiopancreatography [ERCP], re-operation). We encompassed both case reports and case series studies. Studies in non-English language or studies with inadequate data were excluded from the analysis. Available data from eligible studies regarding patients’ demographics, surgical operation, involved artery, size, clinical presentation, diagnosis and treatment of PAs as well as clinical outcome were collected and analysed.

RESULTS

The flow diagram of the search of the literature is shown in Figure 1. We found 97 eligible studies and a sum of 134 patients who presented with at least one symptomatic PA following LC during a 26-year period (1994–2020).[6789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102]

Figure 1

Study selection flowchart

Patient demographics, type of surgery and pre-operative interventions

The majority of patients were women (76 of 131 patients –58%). The mean age of the patients was 50.4 (±15.0) years. Most patients (80%) were aged between 30 and 69 years. Majority of the patients (63%) underwent elective LC due to symptomatic cholelithiasis/chronic cholecystitis or gall bladder polyps, whereas the remaining patients underwent urgent or emergency LC due to acute cholecystitis [Table 1]. Ten patients had undergone pre-operative ERCP and one patient had undergone liver biopsy a few weeks before surgery.

Table 1

Patient demographics and type of surgery

Gender
Available data	131/134 patients (98)
Male	55 (42)
Female	76 (58)
Mean age
Available data	134/134 patients (100)
Total	50.4 (±15.0) years
Male	51.8 (±15.5) years
Female	48.9 (±14.5) years
Age distribution (years old)
Available data	134/134 patients (100)
10-19	2 (2)
20-29	10 (7)
30-39	23 (17)
40-49	26 (19)
50-59	34 (25)
60-69	25 (19)
70-79	12 (9)
80-90	2 (2)
Type of surgery
Available data	82/134 patients (61)
Elective	50 (63)
Urgent/emergency	22 (27)

Clinical presentation and upper gastrointestinal endoscopy findings

In 67% of the patients, PAs became symptomatic during the first 4 weeks following LC, whereas in 81% of the patients, PAs became symptomatic during the first 8 weeks following LC. However, symptomatic PAs have been mentioned up to 3 years following the initial operation. The most common symptom was gastrointestinal (GI) bleeding (74%), followed by abdominal pain, jaundice and/or elevated total/direct bilirubin, intra-abdominal bleeding and fever. GI bleeding was usually the result of haemobilia. The classic Quincke’s triad of haemobilia symptoms (upper abdominal pain, upper GI haemorrhage and jaundice) was present in 32% of patients with PA. Less common symptoms included chest pain, back pain and clinical manifestations of pancreatitis [Table 2].

Table 2

Clinical presentation of patients

Time till clinical presentation
Available data	132/134 patients (99)
0-4 weeks	89 (67)
0-8 weeks	107 (81)
9+ weeks	25 (19)
Symptoms
Available data	132/134 patients (99)
Abdominal pain	81 (61)
Chest pain	2 (2)
Back pain	4 (3)
Intra-abdominal bleeding	32 (24)
GI bleeding	98 (74)
Jaundice and/or elevated total/direct bilirubin	65 (49)
Fever	14 (11)
Pancreatitis	3 (2)

GI: Gastrointestinal

Patients with GI bleeding more often presented with haematemesis (59%) and less often with melaena or haematochezia. Nearly 88% of the patients with GI bleeding underwent upper GI endoscopy, whereas 10% of them had to undergo upper GI endoscopy more than once. Haemobilia was confirmed in 37% of upper GI endoscopies. However, more than half of the upper GI endoscopies did not reveal the source of bleeding. In this case, the most common findings were the presence of blood in the stomach and/or duodenum, other irrelevant findings (e.g., mycotic oesophagitis and erosive bulbitis with no active bleeding) or none. Interestingly, in four patients, upper GI endoscopy revealed extrinsic compression of the duodenum and suspected erosion from the PA [Table 3].

Table 3

Symptoms and upper gastrointestinal endoscopy findings in patients with gastrointestinal bleeding

Symptoms
Total patients with GI bleeding	98/132 patients (74)
Haematemesis	58 (59)
Melaena	39 (40)
Haematochezia	8 (8)
Upper GI endoscopy
Available data	82/98 patients (84)
Yes	72 (88)
No	10 (12)
More than one upper GI endoscopies	8 (10)
Upper GI endoscopy findings
Total number of upper GI endoscopies	85
Available data	84/85 endoscopies (99)
Haemobilia (bleeding from the ampulla of Vater)	31 (37)
Extrinsic compression of the duodenum and suspected erosion	4 (5)
No source of bleeding (blood in the stomach and/or duodenum with no active bleeding, irrelevant or no findings)	49 (58)

GI: Gastrointestinal

Mechanism of injury, concomitant injuries, involved arteries and pseudoaneurysm size

The majority of cases were due to dissection or trocar injury (48 of 73 patients –66%). In 28% of cases, there was a bile duct injury or leak from the cystic duct stump. Nevertheless, in about one-third of the patients, PAs presented following an uneventful LC. The most common involved artery was the right hepatic artery (main trunk or branches –70%), followed by the remnant of cystic artery (19%), whereas in 3% of cases, PAs were located in both right hepatic artery and the remnant of cystic artery. Less commonly, the PAs were located in the abdominal aorta and proper hepatic, left hepatic, gastroduodenal and right renal arteries. The mean diameter of the PAs was 2.7 (±1.6) cm. The range of the maximum diameter was 0.4–9.1 cm [Table 4].

Table 4

Mechanism of injury, concomitant injuries, involved arteries and pseudoaneurysm size

Mechanism of injury
Available data	73/134 patients (54)
Uneventful	25 (34)
Dissection injury	46 (63)
Trocar injury	2 (3)
Concomitant injuries
Total injuries	37/134 patients (28)
Bile duct injury	33 (89)
Leak from the cystic duct stump	4 (11)
Involved artery
Available data	126/134 patients (94)
Right hepatic artery (main trunk or branches)	88 (70)
Remnant of cystic artery	24 (19)
Right hepatic artery and remnant of cystic artery	4 (3)
Proper hepatic artery	3 (2)
Left hepatic artery	2 (2)
Abdominal aorta	2 (2)
Right and left hepatic arteries	1 (1)
Gastroduodenal artery	1 (1)
Right renal artery	1 (1)
Maximum diameter
Available data	74/134 patients (55)
Total number of PAs	76
<1 cm	3 (4)
1-1.99 cm	11 (14)
2-2.99 cm	30 (39)
3-3.99 cm	10 (13)
4-4.99 cm	14 (18)
≥5 cm	8 (11)

PA: Pseudoaneurysm

Treatment, complications and outcome

Only 27% of the patients were treated at the institution where they underwent LC. The rest were treated in another/higher institution. Trans-arterial embolisation (TAE) was the most common first-line treatment (75%), whereas surgical repair (SR) was the second-most common first-line treatment (19%). The success rate of TAE was 83%, whereas the corresponding rate for SR was 64%. Regardless of the first-line treatment, 5% of the patients had at least one recurrence of bleeding [Table 5]. Among patients who underwent TAE, about 10% of them presented post-embolisation syndrome and/or hepatic ischaemia. A relatively rare but notable complication following TAE was the unintentional migration of coils to the adjacent anatomical structures, which caused occlusion of normal blood supply, stone formation, ascending cholangitis, acute pancreatitis, etc.. With a median follow-up of 9 months, the mortality rate was 7%. The PAs of abdominal aorta carried the highest risk of death.

Table 5

First-line treatment

First-line treatment
Available data	134/134 patients (100)
TAE	100 (75)
Covered stent placement	3 (2)
TAE and covered stent placement	1 (1)
Surgical repair	25 (19)
Ultrasound-guided percutaneous injection of thrombin	2 (2)
Endoscopic and percutaneous ultrasound-guided intervention	1 (1)
None	2 (1)
Success rate of first-line treatment
TAE	83/100 (83)
Surgical repair	16/25 (64)

TAE: Trans-arterial embolisation

DISCUSSION

Although LC offers many well-known benefits, it is also associated with higher rates of bile duct and vascular injuries compared to that of conventional open procedure. The incidence of biliary tree injuries following LC is estimated to be up to 1%, whereas the corresponding incidence for vascular injuries is <1%.[28103104] Vascular injuries are more common in the presence of a bile duct injury. It is estimated that vascular injuries co-exist in approximately 25% of patients who suffer a bile duct injury.[5] Vascular injuries during LC mainly occur either during trocar insertion or during dissection of the Calot’s triangle structures. Injuries during trocar insertion are less frequent and affect the abdominal wall vessels or major intra-abdominal vessels (e.g., aorta, iliac vessels and renal vessels). These injuries are mainly caused by the placement of the optical trocar. Secondary trocars are inserted under vision and that contributes in identifying the nature of trocar injury. Injuries during dissection of the Calot’s triangle structures are more common and usually affect the right hepatic artery and/or the cystic artery.[3] Vascular injuries during LC may lead to perioperative bleeding or may present late as hepatic infarction or as PA.[8]

PAs (also known as false aneurysms) occur when an arterial wall is injured, resulting in a rupture and formation of a sac walled by surrounding tissues and organs.[105] Most of the PAs are iatrogenic due to surgical, endoscopic and interventional radiological procedures. Other causes include trauma and infectious and inflammatory conditions (e.g., pancreatitis and cholecystitis).[9106] Previous studies have shown that the incidence of PAs following an LC procedure ranges between 0.06% and 0.6%,[210] while it is much higher (about 4.5%) in patients with a concomitant bile duct injury.[11] However, the actual incidence of PAs following LC is underestimated. Some PAs may be too small to be identified even on imaging, whereas others may be asymptomatic, or may thrombose spontaneously.[6]

The etiopathogenesis of PAs following LC is not clear and may be multifactorial. Injury to the arterial wall resulting in the development of a PA may be caused by the application of excessive mechanical or thermal energy, by metal clip intrusion and by the presence of bile acids or infection.[81112] Excessive dissection or use of electrocoagulation during the dissection of Calot’s triangle may lead to laceration, transection or occlusion of an artery and a concomitant formation of a PA.[12] Thermal energy may be transferred directly by a cautery or indirectly through a metal clip in contact with the artery.[10] Bile acids have complex (amphipathic) solubility properties, and they are powerful solubilisers of membrane lipids.[11] It has been shown that bile acids have fibrolytic and/or cytotoxic effects, and they delay wound healing.[107] Other risk factors for the formation of a PA are the presence of infection, the presence of past or ongoing inflammation, the presence of anatomical variations of the bile ducts and the cystic artery and limited surgical experience.[111314108]

PAs following LC tend to enlarge and rupture. Rupture occurs in 21%–80% of cases, and it is associated with a mortality rate up to 43%.[109] Aneurysms larger than 5 cm are ten times more likely to rupture.[9] Most PAs following LC develop in the right hepatic artery or cystic artery.[6] The time interval between LC and the onset of symptoms is variable. Most cases present during the first weeks following LC, but delayed presentation up to 5 years has been described.[21] Unruptured PAs are usually asymptomatic or may cause abdominal discomfort and pain.[9] Patients with ruptured PAs present with bleeding, either GI bleeding and/or intra-abdominal bleeding (haemoperitoneum).[6] GI bleeding following PA rupture is usually the result of haemobilia. Haemobilia is defined as bleeding into the biliary tree and occurs when a fistula is formed between the splanchnic circulation and the biliary tree. Currently, about two-thirds of haemobilia cases are due to iatrogenic injuries. It is characterised by the Quincke’s triad which consists of abdominal pain, jaundice and upper GI bleeding. However, the classic presentation with Quincke’s triad is present in <40% of haemobilia cases. Abdominal pain is present in 70% of haemobilia cases, jaundice in 60% of cases, melaena in 90% of cases and haematemesis in 60% of cases.[16110] Rarely, haemobilia may cause acute pancreatitis.[17]

Unruptured asymptomatic PAs may be diagnosed incidentally with conventional imaging techniques (e.g., ultrasound, computed tomography [CT] scan and magnetic resonance imaging [MRI]). The diagnosis of symptomatic or ruptured PAs requires a high index of suspicion. The presence of PA should be excluded in every patient with abdominal pain and/or symptoms and signs of bleeding during the first weeks following LC. If a patient presents without symptoms and signs of GI bleeding, an ultrasound and/or CT scan of the abdomen should be performed. Ultrasound can be utilised in the initial evaluation and has the ability to identify suspected PAs, but may not detect small PAs. Doppler ultrasound may show the ‘Yin-Yang’ flow within the pseudoaneurysm itself. Triple-phase contrast-enhanced CT scan or CT angiography is the most reliable non-invasive method to detect PAs. It can provide details about the location and size of the PA, it can diagnose PAs which are even thrombosed and it is also helpful in identifying abnormal vessels and anatomic variations. Ultrasound and CT scan may also reveal a fluid collection or free fluid (blood) in the peritoneal cavity.[111112113] If a patient presents with symptoms and signs of upper GI bleeding, an upper GI endoscopy evaluation is necessary to exclude the more common causes of upper GI bleeding (e.g., peptic ulcer disease and oesophageal varices). As mentioned above, GI bleeding following PA rupture is usually the result of haemobilia. However, only 12% of cases of haemobilia are diagnosed endoscopically. In such cases, a side-view endoscopy will show a blood clot or bleeding from the ampulla of Vater. If endoscopy fails to reveal the source of bleeding, a CT scan should be the next step in evaluation. CT scan may show biliary tree dilation, GI distension or PA per se. Ultrasound may also show biliary tree dilation in 28% of cases of haemobilia. If biliary obstruction co-exists, ERCP with sphincterotomy, balloon dilatation and/or stent placement may be useful in the diagnosis of haemobilia and in the treatment of obstructive jaundice and possible bile leaks. Regardless of the presentation, conventional angiography is considered the gold standard for the diagnosis of PAs. Although conventional angiography is an invasive diagnostic method and requires a trained interventional radiologist, it remains the most accurate diagnostic method with reported sensitivity of 90%, it can determine the exact location and size of PA, it can reveal the source of bleeding and it can be quickly converted from diagnostic to therapeutic procedure.[1517110111114115] MRI, MRI angiography and MRCP may sometimes be helpful. For instance, MRI can distinguish a blood clot from a stone in the case of obstructive jaundice.[116]

TAE is the treatment of choice for PAs with a success rate around 75%–100%. Compared with surgery, TAE is considered a less invasive, safer and more effective option. Furthermore, TAE does not need general anaesthesia; it is a shorter procedure, it is associated with shorter length of stay and it can be repeated in the event of recurrent bleeding. Options for TAE include coils, glue, gel foam (Pfizer, MI, USA), n-butyl cyanoacrylate, polyvinyl alcohol particles and thrombin injection. Coil embolisation is the most frequently used technique. Complications of TAE include migration of coils, inadvertent occlusion of other vessels, reperfusion or rupture of PA and infection. Migration of coils in the common bile duct may lead to obstruction, stone formation, jaundice, ascending cholangitis and even pancreatitis. Inadvertent occlusion of hepatic artery may cause hepatic ischaemia and rarely liver failure. Reperfusion of PA is observed in up to 20%–25% of cases, and it may be due to a long parent artery supplying a large PA with extensive collateral circulation.[117118] Surgery is indicated when TAE fails, if it is not available or contraindicated and when there are associated injuries that require surgical treatment.[18119] Surgical options include either exclusion (ligation of feeding vessels) or excision of the PA. Percutaneous image-guided thrombin injection is another option when TAE fails. Endoscopic ultrasound-guided injection and obliteration of PAs have been used in selected cases.[8]

Limitations

The limitations of this systematic review reflect the limitations of the studies included. The majority of studies were case reports, while nine studies were small case series with 2 to 14 patients each. These studies have been published over a long period of time and originate from institutions with different levels of experience and capabilities. The main limitations were (a) the lack of a unified reporting system of the various clinical and laboratory data we examined and (b) the insufficient data in many studies, mainly regarding the type of surgery, the mechanism of injury, the size of the PAs and the follow-up. All of the above limit the validity of the conclusions.

CONCLUSION

PA formation following LC is a rare but potentially fatal complication. Prompt recognition and treatment are essential. Clinicians should have a high index of suspicion for every patient that present with abdominal pain or unexplained bleeding following LC. Most of the PAs are located in the right hepatic artery or in the remnant of cystic artery, and they can rupture several weeks following the initial operation. Angiography is invaluable for diagnosis in these cases. TAE is the treatment of choice for PAs following LC, while surgery should be a second-line treatment.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.