ERCP in babies: Low risk of post-ERCP pancreatitis – results from a multicentre survey
Note
DGVS: German Society of Gastroenterology (https://www.dgvs.de/wissen-kompakt/forschung-aktuell/studienankuendigungen/)
Abstract
Background and aims
Endoscopic retrograde cholangiopancreatography (ERCP) is rarely performed in newborns, and the risk of post-ERCP pancreatitis (PEP) has not been defined in this age group. We therefore performed a European multicentre analysis of PEP rates and risk factors in children aged ≤1 year.
Patients and methods
Based on a sample size estimation, 135 consecutive ERCPs in 126 children aged ≤1 year were evaluated from five European centres, and the first ERCP per child analysed. All ERCPs and clinical reports were reviewed manually for PEP and associated risk factors. All ERCPs were performed by endoscopists with high ERCP expertise.
Results
No PEP was observed (0/126, 0.0%, CI 0–2.9%) despite the formal presence of multiple risk factors and despite lack of PEP prophylaxis (except one patient having received a pancreatic duct stent). The PEP rate was significantly lower than the PEP rate expected in adults with similar risk factors.
Conclusions
ERCP in children aged ≤1 year is safe in terms of PEP. The PEP risk is significantly lower in children aged ≤1 year than in adults, therefore no PEP prophylaxis seems to be needed in young children. Risk factors from adults may not apply to children under 1 year. Reluctance to perform diagnostic ERCP in suspected biliary anomalies should not be based on presumed PEP risk.
Introduction
Endoscopic retrograde cholangiopancreatography (ERCP) is rarely performed in newborns. In adults, ERCP is the standard of care for therapy of most biliary diseases and the gold standard for imaging of the bile or pancreatic duct; however, diagnostic ERCP has largely been replaced by magnetic resonance cholangiopancreatography (MRCP) because of the risk of post-ERCP pancreatitis (PEP) as its most dismal adverse event.
The ESGE/ESPGHAN guideline suggests that diagnostic ERCP in neonates and infants (≤1 year old) with cholestatic hepatobiliary disease is considered if non-invasive investigations are not conclusive in order to allow timely referral or to avoid unnecessary surgery in suspected biliary atresia1. MRCP does not offer sufficient resolution for diagnosis of biliary disease in newborns. PEP risk factors in adults have been extensively documented,2 but data in children are conflicting: the overall PEP rate was 10.9%; pancreatic duct (PD) injection was the strongest predictor of PEP (odds ratio (OR) 30.8).3 Age > or <10 years was not associated with differences in PEP rate.3 In other paediatric series, focussing on the technical success of ERCP among patients <19 years old,4–7 low PEP rates were reported in newborns despite classical risk factors for PEP. Total patient numbers, however, were too low and follow-up too unstructured for substantiated PEP risk assessment. We therefore performed a multicentre analysis of PEP rates and risk factors in children ≤1 year old.
Patients and methods
There are no prospective data on PEP rates to substantiate a reliable sample size calculation in children ≤1 year old. The following assumptions were made to test the hypothesis that the PEP rate is lower in infants versus adults: the PEP rate in adults is significantly elevated after pancreatic contrast (including parenchymal), contrast-guided (rather than wire-guided) intubation, PD wire passage, with lack of dilated duct, young age, native papilla, and difficult cannulation.2 With alpha <0.05 and a two-sided test, n = 100 ERCPs in babies were required to prove a difference in presumed adult PEP rate of 10% vs. 3% in children (power of 81%).
After sample size calculation, 19 large endoscopy and paediatric centres (mostly in Germany) were contacted, and the trial registered at the DGVS study registry.a Patients were eventually included at the university centres of Essen, Amsterdam, Tübingen, Mannheim and Freiburg. Examinations were collected to January 2019 and dated back to 2009 (two patients were recruited from earlier examinations). Six patients have been included in a previous analysis on effectiveness of ERCP.4 PEP was defined according to the revised Atlanta classification.8 All original ERCP reports were accessed and reviewed manually at the study centre in Tübingen, and the parameters listed in Table 1 extracted, where available. A paediatric duodenoscope (PJF-160 (Olympus, Japan)) was used in 118 patients (93.7%).
| Total (n = 126) | Relative to available information | Information available for n = | |
|---|---|---|---|
| Sex (f/m) | 56/70 | 56/70 | 126 |
| Age [d] (median, IQR) | 64 (47–115.5) | 64 (47–115.5) | 126 |
| Height [cm] (mean ± SD) | 57.2 ± 7.2 | 57.2 ± 7.2 | 115 |
| Weight [g] (mean ± SD) | 4681 ± 1625 | 4681 ± 1625 | 121 |
| Indications | |||
| Extrahepatic biliary atresia | 85(67.46%) | 85(67.5%) | 126 |
| Choledocholithiasis | 17(13.49%) | 17(13.5%) | |
| Choledochal cysts | 12(9.52%) | 12(9.5%) | |
| Unexplained cholestasis | 6(4.76%) | 6(4.8%) | |
| Stenosis | 5(3.97%) | 5(4.0%) | |
| Pancreatic duct leakage | 1(0.79%) | 1(0.8%) | |
| ERCP | |||
| Diagnostic | 108(85.7%) | 108(85.7%) | 126 |
| Therapeutic | 18(14.3%) | 18(14.3%) | |
| Cannulation (method) | |||
| Wire-guided | 10(7.9%) | 10(8.1%) | 124 |
| Contrast-guided | 93(73.8%) | 93(75.0%) | |
| Wire+contrast | 21(16.7%) | 21(16.9%) | |
| No data | 2(1.6%) | ||
| Cannulation (difficulty) | |||
| Uncomplicated | 22(17.4%) | 74 | |
| Difficult | 52(41.3%) | ||
| No data | 52(41.3%) | ||
| Wire passages into PD | |||
| 0 | 112(88.9%) | 112(93.3%) | 120 |
| 1 | 6(4.8%) | 6(5.0%) | |
| 2 | 2(1.6%) | 2(1.7%) | |
| No data | 6(4.8%) | ||
| Pancreatic contrast | |||
| Head | 4(3.17%) | 4(3.3%) | 121 |
| Corpus | 2(1.59%) | 2(1.7%) | |
| Cauda | 31(24.60%) | 31(25.6%) | |
| Parenchymal | 51(40.48%) | 51(42.1%) | |
| None | 33(26.19%) | 33(27.3%) | |
| No data | 5(3.97%) | ||
| Interventions | |||
| Sphincterotomy | 14(11.1%) | 14(11.1%) | 126 |
| Bile duct flushing | 2(1.6%) | 2(1.6%) | |
| Stone extraction (basket or balloon) | 2/3(1.6/2.4%) | 2/3(1.6/2.4%) | |
| Stent (5 CBD, 1 PD) | 6(4.8%) | 6(4.8%) | |
| Serum lipase | |||
| Normal | 31(24.6%) | 31(24.6%) | 126 |
| Elevated (>90 U/l) | 43(34.1%) | 43(34.1%) | |
| Not tested | 52(41.3%) | 52(41.3%) | |
Uncomplicated cannulation was assumed when time to cannulation was less 5 min (noted in the report or seen on fluoroscopy), fewer than five attempts at the papilla, or ‘easy’, ‘immediate’, ‘prompt’, etc. cannulation was noted in the ERCP report. Difficult cannulation was assumed with more than 5 min or five attempts at the papilla according to ESGE recommendations.2 As these data were not included explicitly in most of the reports (and most cannulation attempts in babies are performed without a wire), we tried to gain indirect evidence to define difficult cannulation: if at least two catheters or wires were used, precut sphincterotomy was performed, double wire technique was used, multiple/difficult attempts or manipulations were noted in the report, or the common bile duct (CBD) could not be contrasted. If none of the above-mentioned criteria could be applied, uncomplicated intubation/contrast was assumed. Follow-up was >4 h to 3 days (except for two babies; 1 and 3 h). Follow-up differed due to frequently performed surgery on the day following ERCP in most babies with confirmed biliary atresia. The protocol was approved by the local ethics committee (771/2017BO2).
Results
In 126 infants, 135 ERCPs were performed, and the first intervention per patient analysed. Repeat interventions were not analysed in order to provide per-patient rather than per-examination data. Results are summarized in Table 1. No PEP was observed (0/126, 0.0%, 95%CI 0–2.9%). The predefined statistical endpoint was met with alpha <0.05.
Potential PEP risk factors (as known from trials in adults) included contrast-guided intubation in 73.8% of patients; wire passages into the PD in 6.6%; and pancreatic opacification in 79.8% including parenchymal contrast in 40.5%. Cannulation or contrasting was presumably difficult in 41.3%, and uncomplicated or not specified in 58.7% (17.4% and 41.3%, respectively). No NSAIDs were administered; one patient received a PD stent. All ERCPs were performed by endoscopists with high ERCP experience (>1000 ERCPs lifetime experience). ERCP was diagnostic in 85.7% and therapeutic in 14.3%.
One episode of bleeding was recorded after sphincterotomy that subsided after 3 min of compression, and one episode of contrast extravasation was noted during endoscopy that necessitated termination of the intervention but no additional measures.
Discussion
We did not observe any episodes of PEP in 126 children ≤1 year old (upper confidence interval limit 2.9%), despite multiple risk factors identified in adult PEP trials, that is, pancreatic parenchymal opacification, native papilla, difficult cannulation, lack of bile duct dilatation in most patients, young age; and despite lack of PEP prophylaxis. According to our data, the PEP risk in children ≤1 year old is significantly lower than the approximated risk in adults with similar risk factors.
This low PEP risk may have several explanations. First, mild PEP may have been underreported, as abdominal pain is difficult to measure in newborns, and (asymptomatic) hyperlipasemia was common. However, long-term clinical observation on specialized wards was available for almost all patients. Standardized post-interventional laboratory and imaging follow-up may have resulted in higher detection rates of abnormalities; however, PEP is predominantly a clinical diagnosis.2 Underreporting of PEP would thus be limited to mild episodes not requiring change in clinical management. Similarly, retrospective assessment of the difficulty of papillary intubation may not be as accurately defined as in a prospective trial; however, this does not impact on the clinical PEP diagnosis. Second, ERCPs were performed by experienced endoscopists, and ERCPs were mostly diagnostic. However, no consistent effect of training status on PEP rates has been found in previous trials.2 Third, the neonatal pancreas may be less susceptible to injury by ERCP. Risk factors derived from adult PEP trials may thus not be adequate for newborns.
Our findings have several implications: ERCP in children ≤1 year old seems safe in terms of PEP, and PEP prophylaxis is not routinely necessary in newborns. Reluctance to perform diagnostic ERCP in suspected biliary anomalies (i.e. to replace operative cholangioscopy) should not be based on presumed PEP risk. Future trials to test PEP prophylaxis in newborns will require high patient numbers (in fact, PEP prophylaxis may be worse than none3) and/or will lack adequate power based on our PEP rate of 0.0% [CI 0–2.9%].
It will be interesting to identify which factors (if any) protect the newborn’s pancreas from developing pancreatitis, given the lack of highly effective prophylaxis or therapy in adults. Future trials may concentrate on delineating the age-related physiological water shed towards PEP vulnerability and PEP prophylaxis requirements.
Declaration of conflicting interests
The authors declare that there is no conflict of interest.
Ethics approval
The protocol was approved by the local ethics committee (771/2017BO2).
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Informed consent
Informed consent of patients/parents was not required by the EC
