Risk factors for post-ERCP cholecystitis: a single-center retrospective study

The medical records of patients with gallbladder in situ who underwent ERCP for the first time in the hospital from January 2012 to December 2016 were reviewed and analyzed retrospectively. Patients who had concomitant acute cholecystitis at the time of ERCP or had a previous ERCP history were excluded from the study. The medical records of eligible patients were reviewed retrospectively to identify any occurrence of acute cholecystitis within 2 weeks after ERCP.

The following predefined parameters were analyzed for PEC within 2 weeks. The demographic information included the following: age and sex. The past history included the following: acute pancreatitis, chronic cholecystitis, acute cholangitis, hypertension, hyperlipidemia, and diabetes mellitus. The laboratory examination indexes before ERCP were as follows: alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, gamma-glutamyltranspeptidase, total bilirubin, direct bilirubin, leukocyte count, hemoglobin, and platelet count. The indexes during ERCP were as follows: gallbladder opacification, biliary duct stent, and common bile duct (CBD) diameter. Other factors before ERCP included temperature and antibiotics. During the bile duct opacification, we recorded gallbladder opacification if contrast medium entered into gallbladder and the outline of gallbladder could be seen. No additional efforts were made to get the entire gallbladder outlined by contrast medium if this had not been accomplished simultaneously with bile duct opacification.

Duodenal side-viewing endoscopes (JF-260, TJF-240, or TJF-260; Olympus, Tokyo, Japan) were used to perform the ERCP procedure. The patients were under midazolam sedation. Sphincterotomy was performed using a standard sphincterotome and/or a needle knife. Balloon sphincteroplasty was performed using a Boston Scientific controlled radial expansion balloon with a diameter range of 12–15 mm, 15–18 mm or 18–20 mm). Stones were extracted using retrieval baskets and/or balloon-tipped catheters. An endoscopic mechanical lithotripsy or laser lithotripsy was attempted to crush down the stones if the stones were too big to remove. Obstructive jaundice resulting from malignant bile duct stenosis was treated by placing nasobiliary drainage (ENBD), plastic stents, or self-expandable biliary metal stents. The benign biliary stricture was treated by dilation or placement of plastic stents or fully covered self-expandable biliary metal stents. Pancreaticobiliary maljunction or pancreas divisum was treated by placing ENBD or plastic stents.

Acute cholecystitis was diagnosed according to the 2018 Tokyo guidelines of acute cholecystitis [5]. The diagnostic criteria were based on the following three aspects: (A) local signs of inflammation, including (1) Murphy’s sign and (2) right upper abdominal quadrant mass/pain/tenderness; (B) systemic signs of inflammation, including (1) fever, (2) elevated CRP, and (3) elevated WBC count; and (C) imaging finding characteristic of acute cholecystitis. A definite diagnosis was as follows: one item in A + one item in B + C.

Mean ± standard deviation was used to describe deviation of the data with the normal distribution of the variables. Median (quartile spacing) was used to describe the data that did not meet the normal distribution of the variables. Frequency (percentage) was used to describe the classification of variables. The differences between groups were compared using t-test, chi-square test, or rank-sum test [6]. Logistic regression was used to analyze the findings of a multivariate analysis of acute cholecystitis after ERCP [7]. The nomogram was used to visualize the logistic regression model [8]. The Bonferroni method was used to calibrate the adjusted test level for pairwise comparison of the findings of the chi-square test. Binned Scatterplot was used to describe the relationship between preoperative leukocytes and the risk of acute cholecystitis within 2 weeks after ERCP. SPSS 13.0 was used for statistical analysis. pROC and rms package in R 3.3.3 software were used to construct receiver operating characteristic (ROC) curve and nomogram. A two-tailed value of P <0.05 was established as the threshold of statistical significance.

A total of 4238 patients who underwent the first ERCP procedure between January 1, 2012, and December 31, 2016, in the hospital were included. Of these, 1352 patients were excluded from the study due to concomitant acute cholecystitis (n = 182) or a history of cholecystectomy (n = 1170) before ERCP. Further, 214 patients with more than 15% of missing data were also excluded. Finally, 2672 patients with intact gallbladder were included in the retrospective analysis to analyze the incidence of acute cholecystitis within 2 weeks after the initial ERCP. The mean age of the patients was 62.4 ± 16.2 years (range, 1–106 years); 1166 patients (43.6%) were female (Table 1). Also, 36 patients (incidence rate of 1.35%) finally developed acute cholecystitis within 2 weeks after the first ERCP (Fig. 1).

The results of univariate analysis of potential risk factors for the development of acute cholecystitis within 2 weeks after ERCP are shown in Table 1. The following parameters were found to be closely correlated with PEC in the univariate analysis: history of acute pancreatitis (χ² = 17.754, P < 0.001), chronic cholecystitis (χ² = 20.815, P < 0.001), gallbladder opacification (χ² = 11.816, P = 0.001), bile duct stents (χ² = 15.805, P = 0.001), leukocyte count before ERCP (Z = − 3.610, P < 0.001). The multiple logistic regression analysis identified the following variables significantly correlated with post-ERCP acute cholecystitis (Table 2): history of acute pancreatitis (OR = 2.60; 95% CI: 1.29–5.23; P = 0.007); history of chronic cholecystitis (OR = 8.47; 95% CI: 2.54–28.24; P = 0.001), gallbladder opacification (OR = 2.79; 95% CI: 1.37–5.70; P = 0.005), biliary duct metallic stent placement (OR = 3.66; 95% CI: 1.78–7.54; P<0.001) and leukocyte count before ERCP (OR = 1.10; 95% CI: 1.04–1.17; P = 0.001). In the 36 patients who developed PEC, 29 had gallstones.

The multivariate models were built to predict the incidence of acute cholecystitis after ERCP within 2 weeks. According to the ROC of the multivariate model, the area under the curve (AUC) was 0.852; the sensitivity and specificity were 82.3% and 73.3%, respectively (Fig. 2). The result revealed a good concordance and a good predictive ability.

Finally, the correlation between white blood cell counts before ERCP and PEC was estimated using a binned scatterplot diagram (Fig. 3). The results indicated a curvilinear relationship; also, the risk of PEC increased with the increase in preoperative WBC.