VASCULAR COMPLICATIONS IN PEDIATRIC PANCREATITIS: A CASE SERIES

Abstract

We reviewed INSPPIRE (INternational Study Group of Pediatric Pancreatitis: In Search for a CuRE) database for splanchnic venous thrombosis or arterial pseudoaneurysms to determine the incidence, risk factors and outcomes of peripancreatic vascular complications in children with acute recurrent pancreatitis (ARP) or chronic pancreatitis (CP). Of 410 children with diagnostic imaging studies, vascular complications were reported in 5 (1.2%); 2 had ARP, 3 CP. The vascular events were reported during moderately severe or severe acute pancreatitis (AP) in 4, mild AP in 1. Venous thrombosis occurred in 4, arterial pseudoaneurysm (left gastric artery) in 1. 2 patients with venous thrombosis were treated with anticoagulant, 1 achieved recanalization (splenic vein). In 2 patients who did not receive anticoagulants, 1 re-canalized. No adverse effects were observed with anticoagulants. The child with pseudoaneurysm underwent aneurysmal coiling. Anti-coagulants appear to be safe in children with acute pancreatitis, their long-term benefit needs to be further investigated.

INTRODUCTION

Portosplenomesenteric venous thrombosis (PSMVT) or pseudoaneurysm is present in ~10–20% of adults with acute pancreatitis (AP) or chronic pancreatitis (CP) and usually associated with severe episodes (1, 2). The splenic vein is the most commonly identified vein involved in PSMVT and in some cases, thrombosis may extend to the portal vein, causing portal vein thrombosis (PVT) and extrahepatic portal hypertension (3). The benefit of anticoagulant use is controversial (3, 4) and there is not enough evidence to reach clinical consensus without prospective randomized studies (2).

The experience in pediatrics is limited to a few case reports. In a retrospective study of computerized tomography (CT) scans of 100 adults and adolescents with AP, the incidence of PSMVT was 13–19% and correlated with disease severity (5). The average age of this cohort was 51 years, thus not reflective of a pediatric age group. di Francesco et al. reported 3 children with CP and PVT (6); the study was retrospective, and the authors thought that portal hypertension came first, pancreatitis second. Jancsó et al. reported an 11-year-old girl with CP caused by a novel PRSS1 mutation and prehepatic portal hypertension caused by PVT, but no information was given about interventions or outcome (7). Arterial thrombosis is an uncommon complication of pancreatitis in children with only a few cases reported (8). Overall, the incidence, risk factors and natural history of peripancreatic vascular thrombotic events in childhood are unknown.

In this study, we accessed and leveraged on the INSPPIRE (International Study Group of Pediatric Pancreatitis: In search for a cuRE) cohort to describe the frequency, risk factors and outcomes of peripancreatic vascular complications in children with ARP or CP.

METHODS

Study design and participants

Clinical information was obtained from children with ARP or CP aged ≤ 19 years enrolled at participating INSPPIRE centers using the standard diagnostic criteria (Table S1). Children with the diagnosis of splenic, portal, mesenteric thrombosis and arterial pseudoaneurysms were identified through the review of imaging data by pediatric radiologists at each site.

All data was collected through standardized patient and physician questionnaires entered into REDCap (Research Electronic Data Capture, Vanderbilt University, Nashville, Tennessee, USA), from March 2012 to February 2017. All centers obtained institutional review board approval for this study or the equivalent for their country prior to enrolling subjects.

INSPPIRE database collects information on the severity of the AP attacks including organ failure and local pancreatic complications, but not the duration of organ failure (<48h vs. >48h), therefore it cannot distinguish between moderately severe and severe AP as per NASPGHAN classification of pediatric acute pancreatitis (9). The AP severity was classified as mild and moderately severe or severe.

The primary outcome was re-canalization of thrombosis with or without anticoagulant use. Secondary outcomes were the intensive care unit stay, presence of pancreatic necrosis or pseudocyst, organ failure, development of exocrine pancreatic insufficiency (EPI) and/or diabetes mellitus in children with vascular complications.

RESULTS

Of 441 children in INSPPIRE; 58% had acute recurrent pancreatitis; 42% had chronic pancreatitis. 349 (~ 79%) had more than 1 episode of acute pancreatitis with a total of 972 attacks (~ 3 attacks per subject). 92 (21%) who met the diagnostic criteria for CP at the time of enrollment have reported either 1 or no previous AP attack. Most children had mild attacks of acute pancreatitis (70.5%), moderately severe or severe episodes were rare (1.5%); the severity of attacks were not known in 28%.

Of 441 children, at least 1 imaging study was recorded in 410. Abdominal ultrasound results were not collected at baseline visits, therefore 31 children either had ultrasound done initially and no further imaging afterwards, or had missing data. Of 410 patients in the INSPPIRE cohort, 645 magnetic resonance imaging/magnetic resonance cholangiopancreatography (MRI/MRCP) and 313 computerized tomography (CT) were reported at the time of enrollment or follow-up; 97 transabdominal ultrasounds were reported at follow-up. MRI was the most commonly used imaging modality (92%), followed by CT scan (50%). We recorded ultrasound findings only at follow-up visits in 13% of children. Out of 410 who were imaged at least once throughout their illness, we identified 5 (1.2%) with peripancreatic vascular complications.

The clinical characteristics of children with vascular thrombosis are summarized in Table 1. No children reported underlying risk factors for portal hypertension or thromboembolic events such as use of contraceptives, sickle cell disease, cirrhosis, total parenteral nutrition. No patients reported cigarette smoking; two had passive exposure to tobacco smoke. No prior history of thrombotic events was present and children did not undergo screening for prothombotic disorders. All were tested for lipid profile abnormalities and only 1 was found to have hypertriglyceridemia (20%), in contrast to 22 of 441 children in the entire INSPPIRE cohort (~5%). No children with vascular complications required intensive care unit stay during the corresponding AP attack. None of the children treated with anticoagulants developed bleeding or systemic side effects.

Table 1:

Clinical Characteristics and Risk Factors of Children with Peripancreatic Vascular Complications

Patient	Sex	Age at first AP	Age at thrombosis	Current diagnosis	Severity of AP at thrombosis	Location of thrombus	Diagnostic Imaging	Risk factors	Genetic variants
1	Female	3 years	13 years	ARP	Mild	Splenic vein	MRI/MRCP	Genetic	PRSS1 R122H
2	Female	12 years	14 years	CP	Moderately severe or severe	Portal vein	US with doppler	Genetic Obstructive (duct obstruction with large stones within the pancreatic duct)	PRSS1 R122C and c.40 +13C>T, CFTR (TG) 11–5T/ (TG) 11–7T
3	Female	4 years	9 years	ARP	Moderately severe or severe	Splenic vein and portal vein	CT scan	Idiopathic	Unknown
4	Female	10 years	15 years	CP	Moderately severe or severe	Splenic vein	CT scan	Hypertriglyceridemia	PPARG
5	Female	3 years	6 years	CP	Moderately severe or severe	Left gastric artery	CT scan	Pancreas divisum	Unknown

MRCP: Magnetic Resonance Cholangiopancreatography; US: Ultrasound; CT scan: Computerized Tomography scan; ARP: acute recurrent pancreatitis; CP: chronic pancreatitis

The type of treatments utilized, primary and secondary outcomes are summarized in Table 2. Of 5 patients, 2 received enoxaparin and 1 recanalized; of the other 3 who did not receive anticoagulants, 1 recanalized. 1 child had coiling of her pseudoaneurysm. There were no reports of pancreatic necrosis or ICU stay. Organ failure was observed in 1 child who was treated with enoxaparin and later recanalized. Pseudocyst was seen in 3 of 5 children. Two patients had endoscopic and surgical interventions unrelated to the thrombotic process. At the last follow-up, only 1 child developed new onset diabetes mellitus 5 years after the thrombotic event. One other child had preexistent EPI and diabetes.

Table 2:

Treatment and Outcomes of Children with Peripancreatic Vascular Complications

Patient	Treatment	Primary Outcome	ICU stay	Organ failure	Pancreatic necrosis	Pseudocyst	Other treatments	EPI	Diabetes mellitus
1	None	Re-canalized with formation of varices/collaterals	No	No	No	Yes	No	No	No
2	None	Did not re-canalize	No	No	No	No	ERCP for stone extraction; Lateral pancreaticojejunostomy	No	No
3	Enoxaparin	Re-canalized	No	Yes	No	No	No	No	No
4	Enoxaparin	Did not re-canalize	No	No	No	Yes	No	Yes (preexistent)	Yes (preexistent)
5	None	Did not re-canalize	No	No	No	Yes	Coiling of pseudoaneurysm; ERCP with pancreatic duct sphincterotomy; Lateral pancreaticojejunostomy	Unknown	Yes

EPI: exocrine pancreatic insufficiency; ICU: intensive care unit; ERCP: endoscopic retrograde cholangiopancreatography

DISCUSSION

Utilizing the INSPPIRE cohort, we found a low incidence (1.2%) of peripancreatic vascular thrombotic events in children with ARP or CP. The thrombotic events seemed to correlate with the severity of pancreatic inflammation. The splenic vein was involved most often, and there were no serious or life-threatening complications. This study, for the first time, reports the incidence of peripancreatic vascular thrombosis in a well-defined cohort of children with ARP or CP and emphasizes that thrombotic events are associated with the severity of acute pancreatitis events.

All children identified with PSMVT were females. This observation is consistent with the case series reported by di Francesco et al (6), in which two of three children were female. In contrast, a male predominance (1, 10) is often reported in adult studies. The reason for female preponderance in pediatrics is unknown. This trend needs to be further investigated.

In this report, PSMVT most often developed with an episode of moderately severe or severe AP, which has also been noted in adult literature (1, 11). An increased risk for splenic vein thrombosis has been identified in patients with severe AP, pancreatic necrosis and peripancreatic fluid collections (1). Risk factors frequently observed in the adult population for thrombosis (i.e. smoking, alcohol) are uncommon in pediatric pancreatitis (12). Two of the children who had thrombosis in our cohort also had passive exposure to tobacco smoke which could have contributed to their risk of thrombosis.

In this study, cationic trypsinogen mutations were found in two of the five patients. Similarly, Jancso et al. (7) reported a pediatric case with PVT and a PRSS1 variant. Children with PRSS1 variants have a greater likelihood of developing CP, presenting with CP at a younger age and progressing to CP faster than children without PRSS1 mutations, therefore may be at a higher risk for PSMVT (13, 14).

Other risk factors for pancreatitis in this study included obstructive factors (i.e., pancreatic ductal stones, pancreas divisum) and toxic-metabolic factors such as hypertriglyceridemia. In this study, all patients were tested but only one was found to have hypertriglyceridemia. Although there are no studies linking hypertriglyceridemia-induced pancreatitis to PSMVT, hypertriglyceridemia can be associated with deep vein thrombosis (15). Hypertriglyceridemia seems to be overrepresented as a risk factor for pancreatitis and thrombosis compared to the rest of INSPPIRE cohort. The small sample size did not allow for statistical comparisons. The role of lipid metabolism abnormalities in PSMVT needs to be further investigated.

In mild acute pancreatitis, the initial imaging may be accomplished via transabdominal ultrasonography, with other imaging (CT, MRI) reserved for more complicated cases or tailored to suspected etiology (16). Magnetic resonance imaging/magnetic resonance cholangiopancreatography is undertaken in the diagnostic evaluation of children with ARP or CP to assess pancreas and its ductal anatomy (17) as it is commonly used in the INSPPIRE study. In general, CT is considered the imaging modality of choice to evaluate for peripancreatic complications (5), but some argue that MRI (18) or color duplex ultrasonography (10) may be superior in detecting peripancreatic vascular complications without exposing patients to ionizing radiation. As in other studies (1, 19), the splenic vein was most commonly involved in our cohort.

The use of anticoagulants in peripancreatic vascular complications remains controversial for adults; there are no data in children. Some studies report recanalization with or without anticoagulant use (1, 19), while others report clear benefit with anticoagulants (4). Some do not report complications with systemic anticoagulant use for PSMVT (19); others describe significant bleeding (20). In our study, only two patients received anticoagulants, and one re-canalized. One patient re-canalized without anticoagulant use. There were no systemic complications associated with anticoagulants.

A potential weakness of this study is incomplete genetic testing. Genetic testing is advised with recurrent attacks of pancreatitis or with chronic pancreatitis as genetic mutations have been implicated as the major cause of early-onset acute recurrent and chronic pancreatitis (13). The INSPPIRE study predated the development of NASPGHAN classification for acute pancreatitis, therefore we could not accurately distinguish between moderately severe and severe AP. Nevertheless, most children in the INSPPIRE study reported mild episodes of AP, thrombotic events were rare and associated with moderately severe or severe AP. The data collection was retrospective, and because the INSPPIRE does not include children with only one episode of acute pancreatitis without evidence of CP, the cohort was limited to children with ARP or CP. The vascular complications associated with a single attack of acute pancreatitis deserve further study. Given the small number of patients, it is difficult to suggest the imaging modality of choice and whether anticoagulant use would affect the outcome. Prospective randomized studies are needed to evaluate the most reliable imaging study for children and the impact of anticoagulants on the natural history.

In summary, peripancreatic vascular complications are rare in children with acute recurrent or chronic pancreatitis. They are associated with moderately severe or severe acute pancreatitis and the splenic vein is most commonly involved. Anticoagulants appear to be safe in treating PSMVT during an acute pancreatitis attack, however their overall benefit and their impact on long-term outcomes needs to be further studied.

What is Known/What is New.

What is known:

• Patients with acute pancreatitis (AP), acute recurrent (ARP) or chronic pancreatitis (CP) may develop portosplenomesenteric venous thrombosis (PSMVT) or pseudoaneurysm.

• These thrombotic events occur in ~10–20% of adults with AP or CP and are usually associated with severe acute pancreatitis.

• The pediatric literature is limited to small number of case reports and series.

What is new:

• In a well-defined cohort of children with ARP or CP, the incidence of PSMVT was 1.2% during an acute pancreatitis event.

• PSMVT was typically associated with an episode of moderately severe or severe acute pancreatitis.

• Anti-coagulants appear to be safe in acute pancreatitis, but their long-term benefit needs to be investigated.