Endoscopic treatment of chronic pancreatitis in pediatric population: Long-term efficacy and safety

Key summary

Endoscopic therapy of CP is an effective treatment in adult cohorts.

Introduction

Chronic pancreatitis (CP), a pathological fibroinflammatory condition of the pancreas which develops in individuals with genetic, environmental, and/or other risk factors, used to be diagnosed very rarely in the pediatric population; nevertheless, it has become increasingly recognized recently.^1-3 Irreversible structural changes result in pain syndrome (recurrent episodes or persistent pain) and can progress to exocrine and/or endocrine pancreatic insufficiency.^1,4 Gene mutations (including PRSS1, SPINK1, CFTR and CTRC mutations) and anatomical variations of the pancreatic ducts are reported to be the most typical risk factors in the pediatric population with CP.^1,5,6

If medical therapy of pain related to CP fails, endoscopic and surgical strategies remain alternative therapeutical options.^4,7 Endoscopic therapy of CP is an effective treatment in adult cohorts,^8,9 and growing evidence of high efficacy has been documented also in the pediatric population.^10-13 Due to a higher rate of complications caused by endoscopic retrograde cholangiopancreatography (ERCP) compared with any other gastrointestinal (GI) endoscopy, careful selection of patients for therapeutic ERCP and use of non-invasive tests for diagnostic purposes of pancreatic disorders are essential. ¹¹ A lack of data on ERCP safety in children remains an issue.^11,14

The purpose of this single-centre retrospective study from a prospectively collected database was to analyse safety and long-term efficacy of ERCP procedures performed on a pediatric population with CP in a tertiary referring centre.

Methods

Records on therapeutic ERCP procedures performed on consecutive patients younger than 18 years, who were referred to the Digestive Endoscopy Unit at Catholic University, Fondazione Policlinico “A. Gemelli” between October 1991 and July 2017, were reviewed.

Magnetic resonance cholangiopancreatography (MRCP) has been performed systematically before the first ERCP in children with suspicion of CP since 1998. A total of 87% (33/38) children underwent MRCP prior to the first ERCP. Based on the pancreatographic findings, type of CP was scored according to the classification of Cremer et al. ¹⁵ Therapeutic interventions included sphincterotomy ± dilatation ± plugs/stone extraction ± stenting.

Detailed characteristics on severity (1–10; 10 being the worst) and frequency of abdominal pain (0 – never, 1 – yearly, 2 – monthly, 3 – weekly, 4 – daily, 5 – continuously), use of analgesics (non-opiates, opiates) and number of episodes of acute pancreatitis prior to the first ERCP is provided in Table 1.

OPEN IN VIEWER

Table 1.

Clinical symptoms and use of analgesia prior to the first ERCP and since the last ERCP. AP: acute pancreatitis. (*): statistical significance; p < 0.001.

Severity of pain before the first ERCP				Wilcoxon test
	Median 8.5; IQR 6.0–10.0
Severity of pain since last ERCP
	Median 1.0; IQR 1.0–3.0 (*)
Frequency of pain before the first ERCP				Wilcoxon test
	Median 4.0; IQR 3.0–5.0
Frequency of pain since last ERCP
	Median 0; IQR 0–1.0 (*)
Non-opiates before the first ERCP				McNemar test
	38/38 (100%)
Non-opiates since last ERCP
	13/38 (34%) (*)
Opiates before the first ERCP				McNemar test
	26/38 (68%)
Opiates since last ERCP
	1/38 (3%) (*)
Absolute number of attacks of AP before the first ERCP			since last ERCP
	35/38 (92%)		8/38 (21%) (*)	Fisher t-test
Number of AP		Number of patients	Number of patients
Multiple (≥4)		18/38	3/18 (one episode in all three children)
3		3/38	1/3 (one episode)
2		6/38	1/6 (one episode)
1		8/38	1/8 (one episode)
0		3/38	2/3 (one episode and two episodes)

ERCP was performed with a standard therapeutic duodenoscope with a 4.2 mm working channel (Olympus TJF series, Tokyo, Japan) in all the cases. Endoscopic accessories commonly used for procedures in adult patients were used for ERCP in this cohort. Radiation exposure was kept to the minimum required for a safe procedure, and protection of radiosensitive organs (gonads, thyroid gland) was provided ¹⁶ (Figure 1). All procedures were carried out under general anaesthesia with endotracheal intubation by a certified anaesthesiologist. OPEN IN VIEWER

Data on therapeutic intervention, including sphincterotomy of the major and/or minor papilla, pancreatic duct stricture dilatation, sphincter dilatation, pancreatic plugs/stone extraction and pancreatic stent insertion were interrogated.

All children stayed overnight and adverse events were recorded accordingly.

Results

In total, 39 children with CP received therapeutic ERCP. One patient was lost to follow-up, therefore data from 38 individuals are provided.

Out of 38 subjects, 21 (55%) were girls. Median age at the first ERCP was 13 years (interquartile range (IQR) 10–14.5) and median post-ERCP follow-up has been 7 years (IQR 1–9). Further descriptive statistics are shown in Table 2.

OPEN IN VIEWER

Table 2.

Descriptive statistics of enrolled patients. Depending on tests of normality, data provided in yellow are the correct one.

	Age at onset of symptoms (years)	Age at first ERCP (years)	Interval symptoms (months)	Current age (years)	Follow-up since first ERCP (years)	Follow-up since last ERCP (years)
Median	10.3	13.0	22.5	21.8	10.5	7.0
IQR (interquartile range)	6.0-13.0	10.0–14.5	11.0–40.0	18.0–28.5	6.5–15.0	1.0–9.0
Mean	9.3	11.9	31.2	23.1	11.2	6.9
StD (standard deviation)	4.7	4.0	32.4	7.1	6.5	5.6
Min.–max. (if applicable)	1 month–17.0 years	2.5–17.5

Positive family history of CP (parents and/or grandparents suffering from CP) has been confirmed in 26% individuals (10/38) with positive CFTR gene in 4/38 (11%).

A total of 87% (33/38) children underwent MRCP prior to the first ERCP with results being indicative of CP.

In the pre-ERCP period, 74% (28/38) children suffered from recurrent acute pancreatitis (with or without further episodes of abdominal pain), 13% (5/38) from a single episode of acute pancreatitis, 5% (2/38) had combination of abdominal pain and a single episode of acute pancreatitis, and 8% (3/38) had history of attacks of abdominal pain.

Anatomical variation of the pancreas was present in 16/38 (42%) children. Pancreas divisum was the most common one (Figure 2). A total of 158 ERCPs (median 3; IQR 1–5) were performed in these 38 children; 71% (27/38) individuals underwent ≥2 ERCPs. The majority of patients had CP type IV (18/38; 47%) according to Cremer’s classification ¹⁵ (Table 3). OPEN IN VIEWER

OPEN IN VIEWER

Table 3.

Details regarding anatomy, repeat ERCP and classification according to Cremer.

Pancreatic ducts anatomy
nomally fused ducts	22/38 (58%)
pancreas divisum	9/38 (24%)
dominant Santorini duct	6/38 (16%)
inverted pancreas divisum	1/38 (2%)
Total number of ERCPs	158 in 38 patients
≥2 ERCP	71% (27/38)
median between the first  and second ERCP	7 months (IQR 2–23)
median between the first  and last ERCP	34 months (IQR 10–118)
Cremer classification
type IV	18/38 (47%)
type Ib	14/38 (37%)
type Ia	5/38 (13%)
type IIH	1/38 (3%)

Extracorporeal shock wave lithotripsy (ESWL) was performed in three individuals (3/38; 8%). Stones and/or plugs were removed in at least one of the ERCPs in 25/38 (66%) children. Three patients who received ESWL had calcifications; the remaining 22 patients had plugs or radiotransparent stones extracted.

In our cohort, major pancreatic sphincterotomy was carried out in 47% (18/38), major and minor pancreatic sphincterotomy in 24% (9/38) and minor pancreatic sphincterotomy in 29% (11/38) children (Table 4). One sphincterotomy of the minor papilla was performed using the major-minor papilla rendez-vous technique.

OPEN IN VIEWER

Table 4.

Summary of therapeutic ERCP interventions.

Sphincterotomy		Further cut and/or balloon dilatation
major pancreatic sphincterotomy	27/38 (71%)	5/27 (19%)
biliary sphincterotomy	19/38 (50%)	2/19 (11%)
minor papilla sphincterotomy	20/38 (53%)	9/20 (45%); max. size of dilatation: 6 mm
Plugs / stones removal	25/38 (66%)
Pancreatic duct stricture	11/38 (29%)	Restenting: 6/11 (55%)
dilatation and stenting	5/11 (45%)	3/6 (50%)
stenting	6/11 (55%)	3/6 (50%)
Biliary duct stricture	3/38 (8%)
stenting	3/38 (8%)

Eleven out of 38 patients (29%) had a stricture of pancreatic duct. This required dilatation with a 4–6 mm balloon and stenting in 45% (5/11) or stenting alone in 55% (6/11) (Table 4). Stents were usually inserted for 1–6 months. Cumulative duration of the stenting period which would be longer than 12 months was observed in three cases: 17 months, 45 months and 56 months. Plastic stents were always used with the exception of one patient who received a fully covered pancreatic metal stent (6 mm diameter) in the setting of a clinical trial during further retreatment when the individual was older than 18 years. If there was no stricture of the pancreatic duct or a stricture after previous papillotomy, either a prophylactic 5 french pancreatic stent or a nasopancreatic drain (for 24 h) were always inserted after a pancreatic sphincterotomy.

Biliary stricture was confirmed in three patients (8%; 3/38) and plastic biliary stenting was performed in all three cases. One more plastic biliary stent had to be introduced in the patient who had a metal pancreatic stent inserted and developed acute cholangitis.

Follow-up

Current age of individuals enrolled in our cohort is 23 ± 7 years. Severity and frequency of abdominal pain improved significantly since the last ERCP compared with the pre-ERCP period (p < 0.001); 66% (25/38) of children became pain free. Use of non-opioid and opioid analgesia decreased significantly since the last ERCP, and so did the number of episodes of acute pancreatitis (p < 0.001) (Table 1).

One patient suffers from diabetes mellitus and 37% (14/38) patients have exocrine pancreatic insufficiency. Current weight, height and BMI of male and female patients are recorded in Table 5. BMI of both groups are within normal limits. A trend towards a statistical significance in higher BMI in males compared with females was found; p = 0.07.

OPEN IN VIEWER

Table 5.

Current weight, height and BMI of female and male patients. Trend towards a statistical significance in higher BMI in males was found; p = 0.07. Depending on tests of normality, data provided in yellow are the correct one.

	Weight (kg)		Height (cm)		Current BMI
	Females	Males	Females	Males	Females	Males
Median	55.0	68.0	165.0	175.0	20.3	22.2
IQR	50.0–60.0	62.0–75.0	160.0–168.0	170.0–176.0	19.5–22.0	20.8–24.1
Mean	56.7	68.4	164.3	174.7	21.0	22.3
StD	7.3	10.2	5.9	6.2	2.2	2.8
Min	48.0	46.0	153.0	162.0	18.4	17.5
Max	73.0	86.0	178.0	188.0	27.3	28.4

One patient required subsequent surgery (hepatico-jejunoanastomosis) for refractory bile duct stricture.

In total, 74% (28/38) patients classify their current clinical status as good, 26% (10/38) as fair, and no patient described their clinical status as being poor.

Discussion

CP in children, an increasingly recognized disease, can interfere with the psychophysical development of such an individual significantly.^1,3,7,17 The possibility to influence abdominal pain and episodes of acute pancreatitis, associated with impaired nutrition, need for admissions and need for surgery, is an imperative in patients with CP, especially in the pediatric cohort. Our study looked at the safety and long-term efficacy of therapeutic ERCP in children and adolescents with CP.

In agreement with data published recently,^10,12,18,19 our series confirmed that ERCP in a pediatric population with CP is a safe therapeutic option, having had five complications after 158 therapeutic procedures (3%).

Our study has confirmed data published by Halvorson et al. that therapeutic ERCP can be performed safely and efficaciously by adult gastroenterologists who have been trained in advanced endoscopy. ²⁰ All individuals in our cohort were endoscoped with an adult duodenoscope, and accessories commonly used for ERCP performed in adults were used. This did not decrease technical success and did not increase complication rate. Our conclusions are in agreement with results provided by Yildrim et al. ²¹

Post-ERCP pancreatitis is the most common complication, yet strong data on risk factors and prophylaxis including rectal indomethacin administration and prophylactic pancreatic duct stenting have been identified in the adult population so far.^22,23 In concordance with European Society of Gastrointestinal Endoscopy (ESGE) recommendations, ¹⁶ diclofenac was not used in our study as the majority of patients were younger than 14 years. Two recent studies analysed factors associated with post-ERCP pancreatitis in the pediatric population and both confirmed that pancreatogram, pancreatic sphincterotomy and pancreatic stenting were associated with higher incidence of post-ERCP pancreatitis. Further (prospective) studies on the role of prophylactic pancreatic stenting will have to take place. On the contrary, both studies have, once again, clearly documented, that CP is negatively associated with post-ERCP pancreatitis.^24,25 Together with the high volume of patients treated at our teriary referring centre, this would explain why the incidence of post-ERCP pancreatitis was low in our cohort (1%, 1/158).

Regarding the patient with acute cholangitis, this occurred after a fully covered metal pancreatic duct stent had been inserted (otherwise, plastic stents were used in all other cases). We are aware of the recent study published by Jeong et al. ²⁶ regarding their initial experience with the use of fully covered self-expandable metal stents for recurrent benign pancreatic duct strictures in children. Their case series (with median age 12 years) showed significant improvement of parameters observed and no post-ERCP complication occurred. The age of our patient was 19, therefore diminutive size of the ampullary area cannot explain the increased risk for obstruction of biliary system by the fully covered metal stent, but rather the presence of such a stent itself. Therefore, further studies on fully covered self-expandable metal stents need to be realized in a cohort of pediatric patients and adolescents with CP.

One post-ERCP cholecystitis was recognized in a child with biliary stricture which had required biliary stenting. The other child with cholecystitis had undergone biliary and pancreatic sphincterotomy of the major papilla without any further interventions, and no risk factors for cholecystitis were noted. Both patients received antibiotic treatment. A conservative approach was successful in both cases.

The only post-ERCP bleeding which required subsequent endoscopy was identified in a 13.5-year-old girl, who underwent sphincterotomy of the minor papilla.

In general, a very important prerequisite is to observe the pediatric patients after the ERCP procedure for 24 hours closely, so that a potential complication can be recognized and treated accordingly. Delayed complications after discharge could not be formally excluded in our study, but to our knowledge none were recorded.

Obstruction of the pancreatic duct was treated by appropriate sphincterotomy in all enrolled patients, by removal of plugs/stones if present and by dilatation and/or stenting of the pancreatic duct stricture if appropriate. This led to significant improvement of pain, significant decrease of episodes of acute pancreatitis and significant decrease in the need for analgesia. The first two outcomes influenced quality of life, undoubtedly. The third outcome is important not only when acute side effects are taken into account (mainly gastrointestinal toxicity of NSAID, status of consciousness during treatment with opiates), but from the long-term point of view, too (nephrotoxicity caused by NSAID medication). Similar to Oracz et al., ²⁷ our study confirmed the efficiency of pancreatic duct stenting with plastic stents in a pediatric cohort with CP. Another important goal was also achieved: after the endoscopic interventions, BMI reached normal ranges in girls and boys.

Typically (almost 80%) children with CP suffer from multiple episodes of mild to moderate abdominal pain, and pancreatic insufficiency is rather rare.^4,28,29 On the contrary, more than 50% adults with a diagnosis of non-alcoholic CP develop a painless form of the disease which is accompanied by a progressive deterioration of exocrine and endocrine pancreatic function. ³⁰ In our cohort, after a median follow-up of 7 years, 37% patients have exocrine pancreatic insufficiency and one patient has diabetes mellitus. It would be interesting to assess the impact of an early endoscopic drainage of the pancreatic duct on endocrine and exocrine pancreatic function long term (median follow-up 10 years). The patient with diabetes mellitus requires insulin treatment. Surprisingly, more than half of patients received opiates before the first ERCP, resulting in an overtreatment aiming at a complete and durable pain relief in this pediatric population.

The main advantage of ERCP over surgical therapy in this setting is that the treatment can be repeated, is less invasive and can be performed even in those with portal hypertension. ⁴ Although 71% of children in our cohort had to undergo ERCP repeatedly, it did not impair their school attendance (admissions were usually short, lasting 2–3 days with prompt recovery) or free time activities including sport. Achievement of ERCP outcomes (mainly with regards to pain) can also indicate that the child would benefit from surgery as a more definitive option. ⁷ Data from large cohorts, which would compare primary endoscopic management with a primary surgical approach of CP in children, are not available. Iqbal et al. enrolled 37 children with CP: 12 children were managed endoscopically, five patients with surgery, and 20 children with surgery after previously failed ERCP. They found that the rate of recurrent pancreatitis was significantly higher in the ERCP alone group compared with patients managed operatively. Yet they did not confirm a difference in frequency of hospitalization required for CP, and narcotic use was not different between these two groups. ³¹ Our study demonstrated a significant drop in episodes of acute pancreatitis and a significant decrease in narcotic use after endoscopic interventions. Surprisingly, the study by Iqbal et al. also reported the same procedure-related morbidity in the ERCP and operative group. ³¹ This was due to a high incidence of post-ERCP pancreatitis (26%) in the ERCP alone group, which was not replicated in our study (1/158; 1%).

In agreement with Sacco Casamassima et al. and recently published guidelines,^32,33 we recommend the proposed step-up practice: only if conservative management, including endoscopic interventions, fails, tailored surgery should follow in children with CP. Organ preservation is the goal; nevertheless, total pancreatectomy with islet autotransplant can be contemplated in carefully selected cases.^33,34

The main limitation of the study is its retrospective and single-centre design.

Based on our data on long-term efficacy and safety, we would recommend endoscopic management of CP in pediatric patients before surgery is contemplated. Still, clinicians should decide carefully whether each child can benefit from the therapeutic ERCP long term, and information should be provided to the family members accordingly.