Risk Factors for post-Endoscopic Retrograde Cholangiopancreatography Abdominal Pain in Patients without post-ERCP Pancreatitis

Background: Most of the studies on the abdominal pain associated with endoscopic retrograde cholangiopancreatography (ERCP) are aimed at solving the pain during the procedure. Post-ERCP abdominal pain has rarely been investigated (cid:0) and few studies have focused on the characteristics and risk factors of post-ERCP abdominal pain without post-endoscopic retrograde cholangiopancreatography pancreatitis (PEP).This study aimed to identify risk factors of post-ERCP abdominal pain without PEP and investigate characteristics of the abdominal pain in non-PEP patients. Methods: From August 6th, 2019, to January 15th, 2020, data of patients who underwent ERCP were retrospectively collected. The characteristics of the abdominal pain after ERCP were recorded and compared between PEP and non-PEP patients. Multivariate analysis was conducted to identify risk factors of non-PEP abdominal pain. Results: Data from 616 ERCP procedures were retrospectively investigated in this study, among which 51(8.28%) patients presented post-ERCP abdominal pain without PEP and 45 (7.31%) patients developed PEP. Multivariate analysis found that 5 risk factors were associated with non-PEP abdominal pain: female gender (OR:2.137), upper abdominal surgery history (OR:1.948), rst time ERCP (OR:4.735), elevated serum γ-glutamyl transferase ( 𝛾 -GT) (OR:2.570) and elevated serum direct bilirubin (DBIL) (OR:2.932). Visual analogue sale (VAS) score of PEP abdominal pain was higher than that of non-PEP pain group (P=0.05). There were no signicant differences in the time of the onset of the pain, pain relief time, pain frequency, use of analgesic medicine, hospital stay and mortality between PEP and non-PEP pain group(P<0.05). Conclusion: This study indicated that female gender, upper abdominal surgery history, rst time ERCP, elevated 𝛾 -GT and elevated DBIL were independent risk factors for post-ERCP abdominal pain without PEP. Abdominal pain was severer in PEP patients than non-PEP patients. PEP:Post-ERCP VAS:Visual analogue EPBD:Endoscopic papillary balloon papillary large-balloon


Introduction
Abdominal pain is often observed in patients after endoscopic retrograde cholangiopancreatography (ERCP) procedure, which sometimes indicates the occurrence of complications such as post-ERCP pancreatitis, perforation, or cholangitis. The most common complication with post procedure abdominal pain is acute pancreatitis, which occurs in 3.5%-14.7% of the cases [1,2] and leads to death in 0.1-1.1% of the patients [2,3] . For clinicians, fear of severe complication with abdominal pain may lead to anxiety and excessive treatment. Moreover, post-ERCP pain may in uence the satisfaction and mental state of patients.
Most of the researches on the abdominal pain associated with ERCP are aimed at solving the pain during the procedure [4] . Post-ERCP abdominal pain is one of the major symptoms of PEP; however, abdominal pain after the procedure also indicates potential occurrence of post-ERCP cholangitis, perforation and nonspeci c etiology ( atulence, gastric and intestinal spasm, etc.). Few studies have focused on the characteristics and risk factors of post-ERCP abdominal pain without PEP.
Therefore, this study aimed to identify the characteristics and risk factors of post-ERCP abdominal pain without PEP. Differences between abdominal pain with or without PEP were compared so as to guide clinical diagnosis and decision-making.

Study design and patients
In this retrospective study, all patients who underwent ERCP from August 6th, 2019, to January 15th, 2020, at Nanjing Medical University A liated Drum Tower Clinical Medical College were included.
Exclusion criteria included any of the following: 1) failed operation; 2) missing any of the indicators. The Medical Ethics Committee of Nanjing Medical University A liated Drum Tower Clinical Medical College approved this retrospective data only study with waiver of informed consent. The need for individual consent was waived for retrospective cohorts.

Data collection
The following data were extracted from the medical charts: patient-related (sex, age, surgical history, gastrectomy history, cholecystectomy history, drinking history, smoking history, hypertension, diabetes, coronary heart disease, chronic pancreatitis history, acute pancreatitis history, white blood cells(WBC), hemoglobin(Hb), platelet(PLT), neutrophil percentage(Neut%), alanine aminotransferase (ALT), aspartate aminotransferase (AST), γ-glutamyl transferase ( -GT), total bilirubin(TBIL), direct bilirubin (DBIL), alkaline phosphatase (ALP), albumin (ALB), prothrombin time (PT), history of pancreatic diseases, suspection of sphincter of Oddi dysfunction (SOD), hilar bile duct stricture, distal biliary stricture, diagnosis of common bile duct stone, and adenoma of the duodenal papilla); procedure-related (pre-cut, endoscopic sphincterotomy, pancreatic guidewire passages, biliary stent placement, endoscopic papillary balloon dilatation(EPBD), endoscopic papillary large-balloon dilation(EPLBD), nasobiliary drainage, di cult biliary cannulation (cannulation attempts of duration > 10 minutes, and/or > 5 attempts), contrast injection to the pancreatic duct, dilated extrahepatic bile duct, type of papillary ori ce, pancreatic duct stenting and operator experience (high grade: >200 ERCP procedures total and/or > 50/year); postoperative outcomes(Visual analogue scale(VAS) score of post-ERCP abdominal pain, time of the onset of post-ERCP pain, pain relief time, pain frequency, use of analgesic medicine, use of protease inhibitors, total hospital stay, hospital stay after ERCP. Indomethacin suppository was given routinely as per the judgment of the operator. If the patient had new or aggravated epigastric pain within 72 hours after procedure, abdominal imaging was routinely performed according to the general situation of the patient, and, if necessary, protease inhibitors or analgesic medicine was given. PEP is diagnosed with 2 out of 3 diagnostic criteria de ned by the Revised Atlanta Classi cation (pancreatic abdominal pain, three times upper limit of normal elevation of serum lipase, imaging suggestive of pancreatitis). Post-ERCP cholangitis is de ned as new onset of fever (> 38℃, lasting over 24 hours) and cholestasis. Post-ERCP perforation is de ned with imaging evidence of gas or luminal contents outside of the gastrointestinal tract. Post-ERCP bleeding is diagnosed with hematemesis and/or melena or hemoglobin drop > 2 g/dL or bloody uid from nasobiliary drainage. Severity grading is conformed with ESGE guideline 2019 [5] . Post-ERCP abdominal pain is diagnosed as a new or aggravated upper abdominal pain within 72 hours after ERCP.

Statistical analysis
The continuous data were dichotomized according to the cut-off value of each indicator based on clinical signi cance or normal range. Categorical data were expressed as numbers, rates, and percentages and were compared with the chi-square test or Fisher's test. The relevant factors found by univariate analyses (P < 0.15) were included in the multivariate logistic regression analysis (forward stepwise regression selection). A receiver operating characteristics (ROC) curve was used to examine the sensitivity and speci city, and the area under the ROC curve (AUC) was used for discrimination.
Propensity score matching was performed to compare the realistic difference between PEP group and non-PEP group with post-ERCP abdominal pain. Prior to propensity matching, t tests or Chi-Square tests were used to assess for differences between groups. PEP group and non-PEP pain group were matched 1:1 on the following variables: age, gender, comorbidities (coronary heart disease, diabetes, hypertension, chronic liver disease, malignant tumors or cancer, chronic pancreatitis), surgery history, operation time, laboratory tests before operation (complete blood count, liver function test, kidney function test, prothrombin time (PT), CA199). McNemar's test was used for binary variables. And a paired Student's ttest or paired-sample test was used for continuous variables. Statistical signi cance was treated as a pvalue < 0.05. All data were analyzed using SPSS 25.

Results
Characteristics of the patients Data of 616 patients who underwent ERCP from August 6, 2019, to January 15, 2020 were retrospectively collected. Among these patients, 45 developed post-ERCP pancreatitis, and 51 presented post-ERCP abdominal pain without PEP, including 11 post-ERCP cholangitis, 1 mild bleeding and 39 nonspeci c abdominal pain. The rest 520 patients did not have any abdominal pain after operation. Patients' characteristics between normal group and non-PEP pain group were presented in Table 1.

Univariate analysis
In the univariate analysis, 14 of 46 patient-related factors and 9 of 23 procedure-related factors were found to be associated with an increased risk of post-ERCP abdominal pain without PEP (Table 1).

Multivariate analysis
Risk factors that reached a p-value cut-off 0.15 or less were used to create the multivariate model.
Variables which were considered to be the risk factors according to clinical experience and previous researches were also analyzed. The identi ed factors were entered into multivariate model ( Table 2). Five risk factors were signi cant by multivariate analysis, three were characteristics of the patients (female gender, rst time ERCP, upper abdominal surgery history), and two were preoperative laboratory examinations (elevated serum -GT and DBIL before ERCP) ( Figure 2). The Hosmer-Lemeshow goodness of t test was not signi cant(p=0.982) at the 0.05 signi cance level and the C-statistic was 0.764. Propensity matching between PEP group and non-PEP pain group In order to identify the characteristics of the post-ERCP pain without PEP, PEP group and non-PEP pain group were compared and analysed. Before propensity matching, there were differences between the PEP group and non-PEP pain group in baseline characteristics. With the use of propensity matching, 34 patients with PEP were matched with 34 patients with non-PEP abdominal pain. The baseline characteristics were comparable after propensity matching (Table 3). Characteristics between two groups were similar in the time of the onset of the pain, pain relief time and pain frequency(P<0.05) ( Table.4). It showed that pain intensity expressed as VAS score of post-ERCP abdominal pain related to PEP was higher than that of non-PEP pain. (P=0.05). There were no signi cant differences in the use of analgesic medicine, hospital stay and mortality between the two groups.

Discussion
Abdominal pain after ERCP is common and is a source of distress to both patient and physician.
Occasionally it predicts complications related to the procedure, especially PEP. However, apart from PEP, some postprocedural pain is of other etiology: 1. Post-ERCP cholangitis, typically presenting with fever, jaundice, and abdominal pain, needing use of anti-biotics or repeated ERCP; 2. Perforation, manifested as severe abdominal and back pain, abdominal tenderness and fever, needing intensive management, even surgery; 3. Nonspeci c pain, often resolving in several hours. Several reports considered nonspeci c pain may in part relate to bowel insu ation with air [4,6] . According to typical clinical symptoms, it is relatively easier to identify perforation, bleeding and cholangitis. Therefore, it is a challenge for early differentiating between the pain of PEP and nonspeci c pain. Empirical data on post-ERCP abdominal pain can be used to optimize the information for patients undergoing ERCP. It may contribute to improvement of a patient's tolerance and the practice guidelines of support treatment, because such data enable better preparation of both patients and physician for the procedure.
In our study, 51 non-PEP pain patients consist of 11 mild post-ERCP cholangitis, 1 mild bleeding and 39 nonspeci c pain. Our results showed that non-PEP pain was associated with female gender, rst time ERCP and upper-abdominal surgery history, as well as patient's elevated pre-ERCP serum -GT and DBIL level. Female gender was found to be the risk factor for post-ERCP abdominal pain without PEP in this study. Tom B. Glomsaker [7] found that female patients demonstrated an increased risk for reporting pain both during and after the ERCP procedure. Though this study did not exclude patients with PEP, this nding is partly in agreement with our result. Patients who underwent ERCP for the rst time have a higher risk of abdominal pain after ERCP. Essink-Bot et al. [8] found that patients who had undergone previous surveillance endoscopies for Barrett's esophagus experienced less discomfort, pain, and overall burden than patients who underwent upper endoscopy for the rst time. They suggested that these previous surveillance endoscopies had resulted in an adaptation to the procedure. Upper-abdominal surgery history may cause postoperative abdominal adhesions and alter anatomy which would increase operative di culty and probability of postoperative abdominal pain. Karayiannakis et al. [9] reported that adhesions in patients with upper-abdominal surgery occurred more frequently, and were more extensive and denser than those in patients with lower-abdominal surgery. It explains why no signi cant difference in the history of lower-abdominal surgery or the history of all types of surgery between groups was found in our study. Moreover, our study found that elevated serum -GT and DBIL before the procedure were risk factors for post-ERCP pain. In contrast, elevated serum DBIL was proved to be protective factors for PEP [10] , which means that if patients with abnormal DBIL present abdominal pain after ERCP, it is more likely to be other etiology rather than PEP.
Most clinicians would agree that patient satisfaction is of great importance. Patient satisfaction related to the ERCP procedure is related to the patients' experiences of pain during and after the procedure regularly [7] . Nonspeci c pain after ERCP may be confused with early symptom of complications, which in uences clinical treatment and causes patients' postoperative anxiety. After propensity-score matching, we found that the pain intensity related to PEP was more severe than that without PEP (4.50 ± 1.36 vs. 3.68 ± 1.81, p = 0.050). Our result was consistent with that of the previous studies. JSGE guideline for post-ERCP pancreatitis [11] showed that strong abdominal pain indicates a high probability of PEP. Hauser G er al. [12] found a threshold VAS score of 5 was strongly associated with the occurrence of PEP. They suggested that the VAS scores for abdominal pain appear to have an excellent diagnostic value for predicting both the presence and absence of PEP. Thus, VAS score may help distinguishing these two types of pain. For this study, it was reasonable to found protease inhibitors were used more frequently in PEP cases. The use of analgesic medicine was similar between PEP group and non-PEP pain group. Clinicians tended to give analgesic medicine rather than protease inhibitors to relief pain rstly in our study. It indicated that non-PEP pain especially nonspeci c pain may lead to analgesic abuse.
Our study has limitations. This is a retrospective, observational study and hence suffers from potential selection and ascertainment bias despite propensity-score matching. Larger sample sizes or prospective studies are necessary to reduce biases.

Conclusion
In conclusion, our study indicated that female gender, rst time ERCP, elevated serum -GT, elevated serum DBIL and upper abdominal surgery history were independent risk factors for post-ERCP abdominal pain without PEP. Characteristics of post-ERCP pain were different in VAS score between PEP patients and non-PEP pain patients. These ndings may contribute to the relief of patients' discomfort and dissatisfaction and therefore improve the quality of ERCP procedures and patient management. University Medical School approved the waiver of consent.

Consent for publication
Not applicable.

Availability of data and materials
The data set analyzed in the current study cannot be opened to public because patients' privacy must be protected and IRB does not permit to do so. However, data are available from the author upon reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was supported by National Natural Science Foundation of China (No.81871947) Authors' contributions XZ and LW designed the research and the study concept; MC and RZ contributed to analysis and interpretation of data; JC and YY contributed to the acquisition of data. MC and RZ wrote the paper. XZ and LW supervised the study; all of the co-authors conducted a critical revision of the manuscript for important intellectual content. All authors read and approved the nal version of the manuscript.  ROC curve for risk of post-ERCP pain without PEP