Transpancreatic biliary sphincterotomy for biliary access is safe also on a long-term scale

The study design was a retrospective case–control study with a phone survey as a cut point. Calculation of the sample size involved estimated PEP rates of 2.2% and 10.4% in regular wire-guided cannulation and in TPBS, respectively (power = 0.8, α = 0.05). According to this calculation, we aimed to both include a minimum of 137 participants in both groups and have a long-term follow-up period lasting at least four years after the first ERCP.

Table 1 presents patient- and ERCP-related characteristics. The majority of the procedures were performed due to CBD stones. Other indications were biliary injury after cholecystectomy, biliary pancreatitis, stricture after liver transplantation and suspicion of SOD. Rare indications in TPBS included three patients with suspicion of papillary tumor or stricture, one patient with non-specific biliary dilatation, one with Mirizzi syndrome, one with choledochal cyst and one with non-specific icterus. Other rarer indications in CG included three patients with non-specified biliary dilatation, one patient with suspicion of gallbladder fistula, one with recurrent AP and one with suspicion of a liver cyst.

The primary cannulation method in the TPBS group was an NK precut in two cases (1.4%) due to a stone in the papilla. After precutting, access was possible only to PD, thus requiring further TPBS. Cannulation methods in TPBS are presented in Fig. 1 and in CG in Fig. 2.

The majority of procedures, that is 125 (87.4%) of TPBS and 131 (93.6%) of CG, were classified as difficulty grade 2 (20).

Table 2 lists the complications related to ERCP procedures. PEP rates were 4.9% in TPBS and 0.7% in CG (p = 0.067, OR for TPBS 7.15, 95% CI for OR 0.869–58.93). Complications related to anesthesia or patients’ comorbidities were pneumonia, acute transient brain ischemic attack and worsening of chronic liver dysfunction. One patient in both groups suffered from late-onset AP, both of whom were classified as mild.

Table 3 presents the PEP rate related to known PEP-risk factors and incidence of PEP. According to these factors, no significant difference existed between the groups. One male and six female patients in the TPBS group suffered from PEP. None of the previously known risk factors for PEP [12, 18, 19] were significant in this study (data not shown).

CP was diagnosed only in one patient (0.7%) in CG and none in TPBS, (p = 0.495). Table 4 presents patient characteristics and phone survey data. The median follow-up period was six years in TPBS and seven years in CG.

In TPBS, 30 patients (21.1%) and in CG 22 patients (15.7%) had taken some medication for abdominal pain or symptoms within the previous year. The most common medications used were PPIs, metamizole-pitofenone, paracetamol and NSAIDs (non-steroidal anti-inflammatory drugs).

Our evaluation pertained to abdominal pain-related contacts to hospitals after index ERCP. Re-ERCP data are shown in Table 5. In all cases, biliary stones were the most common indication of re-ERCPs; however, two cases in TPBS and one case in CG concerned residual stones. Recurrent stones with gallbladder in situ occurred in three cases in both groups. Also application, exchange or removal of biliary or pancreatic stents were a common re-ERCP indication. The group “other” included six ERCPs in CG; one patient underwent four different re-ERCPs due to post-pancreatoduodenectomy stricture. Two other re-ERCPs were due to icterus (suspicion of papillary tumor) in one case and recurrent APs (and later diagnosed as CP) in one case.

We aimed to compare short- and long-term complications between TPBS and CG with native papilla and similar baseline characteristics. The follow-up period was median of 6 years in TPBS and 7 years in CG. All 283 patients were interviewed by phone, with a high participation rate: 84.1% in TPBS and 82.6% in CG. Our study provides a long follow-up period and is thoroughly performed with a matched control group and phone call survey. TPBS had acceptable cannulation success rate of 97.2%.

The criteria of difficult CBD cannulation by the Scandinavian Association for Digestive Endoscopy (SADE) research group and later by ESGE was published in 2014 and 2016 [1, 2]. The patients in our study underwent ERCPs between 2007 and 2013, when the criteria for difficult cannulation were unclear. However, patients in this study needing TPBS for successful CBD cannulation fulfilled the requirements of difficult cannulation in a retrospective evaluation. Fulfilling the requirements mostly related to PD passages, as seen in Table 1. In addition to guidewire passages to PD, the median length of the procedure was significantly longer in TPBS group, (27 min), twice as long as in CG group (15 min), indicating the difficult cannulation situation. There were no differences in performed procedures after cannulation, thus the longer procedure time was related to more time spent to cannulate. It is impossible, however, to determine the exact cannulation times and attempts retrospectively.

Only few studies have evaluated the long-term complication risks of TPBS [22, 23] or pancreatic sphincterotomy (PS) when PS was performed to gain the access to PD [24, 25]. As stated in a recent meta-analysis, long-term follow-up studies in TPBS are lacking [13]. In these studies, the median follow-up period ranged from four to five months in TPBS and 41 months in PS. Results mainly concentrate on AP or other acute complications [5, 22‐26]. However, these studies have several limitations. Miao et al. followed 36 patients with TPBS for four months [22]. During that period, patients had no recurrent biliary stones or symptoms related to SOD. Kahaleh et al. evaluated 116 TPBS patients in a prospective study with a follow-up median of five months but results concentrated on short-term complications [23]. In a retrospective study, Park el al. included only SOD patients with both BS and PS with median follow-up of 41 months, the long-term outcome being a need for reintervention [25]. Kozarek et al. included only patients with CP and PS, 14% of whom required repeated endoscopic or surgical sphincter section [24]. These studies are not fully comparable to ours due to the shorter follow-up period and different patient group. Patients undergoing ERCP with symptoms related to SOD or CP are a different group than our patients who underwent ERCP for benign indications and targeting to CBD. PEP rate in this meta-analysis was 8.1% in prospective studies [13], and it is comparable to our PEP rate of 4.9%.

We have performed TPBS in difficult cannulation situation for several years and collected the data to provide a large patient series to confront the long-term complication issue. We wanted to explore if patients with TPBS have more CP or symptoms related to papillary stenosis, for example AP episodes. Therefore, we needed the follow-up period to last for several years. However, we found no relations between TPBS and CP or papillary stenosis. Concerns with TPBS included increased risk of complications, additionally long-term consequences of PS in cases when the target is not PD, and a risk of papillary stenosis that may lead to recurrent pancreatitis [27‐29].

Late complications like papillary stenosis and ductal strictures are only encountered at least 3 months after the procedure indicating a need for long term follow-up [30].

In difficult cannulation, if advanced cannulation methods, such as TPBS, are needed, the risk of PEP arises [1, 2]. By the ESGE [1], TPBS should be considered as a rescue method in difficult cannulation but performed by experienced endoscopists due to its higher complication rates. Training of the TPBS technique should be focused only for trainees already being able to practice standard ERCP independently. Testoni et al., in their algorithm of biliary cannulation, placed TPBS as a third or a fourth cannulation method after DGW and/or NK precut techniques [31]. Similarly in ESGE guidelines, TPBS is recommended in difficult cannulation cases after failed attempts with standard or DGW cannulation. Sugiyama et al. performed a randomized study and suggested that TPBS should be considered as a rescue method and performed earlier in difficult cannulation [32]. As shown with our data, TPBS is safe also on a long-term scale. Therefore, we also suggest earlier performance of TPBS in difficult cannulation.

In HUS, prophylactic pancreatic stents were seldom used during the study period. The ESGE has suggested that placement of prophylactic stents might reduce the PEP rate [1, 12]. Six patients (4.2%) in TPBS had prophylactic pancreatic stent in our study (Table 1). Still, in this study, a PEP rate of 4.9% in the whole TPBS group is acceptable. At the moment, there are no clear data that after proper PS prophylactic pancreatic stent is beneficial (12) but randomized trials are needed.

Our follow-up study found no patients with stricture or CP in TPBS group. Notable, however, are limitations regarding patient selection and retrospective manner. Patients in both groups suffered from upper abdominal pain. However, no differences existed in the frequency or severity of the pain. Most of the patients had symptoms related to recurrent biliary stones or heartburn. Painkillers or PPIs were equally used in both groups. The most common reason for evaluating the follow-up contacts to healthcare were upper abdominal pain and symptoms related to biliary stones. Due to these symptoms, the most common procedures were cholecystectomy and gastroscopy.

According to a previous study in HUS, the TPBS method has been successful and is therefore commonly used as a rescue method in our unit [7]. We suggest further prospective studies to evaluate the long-term risks of TPBS.

Possible bias in this study relates to the patient selection process. We wanted to interview all the patients; therefore, the inclusion criteria considered only benign indications for ERCP. PEP rates are slightly lower in the present study than in previous studies in the HUS endoscopy unit in ERCP patients with TPBS [7, 33]. Nevertheless, malignancy is not considered a risk factor for PEP. The ESGE guidelines for indomethacin or diclofenac as a PEP prophylaxis [12] was adopted in our unit in November 2013. In this study, only four patients underwent ERCP after November 2013 and only one patient in TPBS group received prophylactic rectal diclofenac. The ESGE definition for difficult cannulation was published in year 2016 [1]. Our study included patients between years 2007 and 2013, and we started to collect the data (cannulation duration and attempts) in March 2011. However, the missing data should not affect at least the long-term complications of TPBS.

TPBS is a useful rescue method in difficult cannulation, having an acceptable complication rate. No significant difference occurred between groups when evaluating the short- or long-term complications. Additionally, no significant differences occurred during the follow-up period concerning upper abdominal pain, usage of painkillers, episodes of AP or development of CP.