Incidence of Cholangitis in Post‐Endoscopic Retrograde Cholangiopancreatography (ERCP): A Systematic Review and Meta‐Analysis

ABSTRACT

Background

Endoscopic retrograde cholangiopancreatography (ERCP) is a procedure that combines x‐ray and upper gastrointestinal (GI) endoscopy for diagnosis and treatment of pancreatic and biliary diseases. Cholangitis is the main infectious complication after ERCP surgery. Current evidence synthesis remains limited by the absence of cholangitis‐specific incidence quantification in prior meta‐analyses, which predominantly report composite infection metrics. This important evidence gap impedes accurate risk assessment for this procedure‐dominant complication. To address this deficit, we conducted a systematic review and meta‐analysis to determine the isolated incidence of post‐ERCP cholangitis.

Methods

We conducted a systematic review and a meta‐analysis of the incidence of cholangitis in Post‐ERCP for studies that were published in both PubMed and EMBASE up to February 6th, 2023. Overall and subgroup pooled effect estimates (continents, age, and publication year) with 95% confidence interval were evaluated by using inverse‐variance pooling and logit‐transformed proportions model. Risk of bias assessment was accessed by using ROBINS‐I guidelines.

Results

Our search initially retrieved 1622 unique studies, of which 97 studies met the inclusion criteria. Overall pooled rate was 1.70% (95% CI: 1.41%–2.04%). The subgroup continent pooled rate for South America, North America, Europe, Asia, and Africa were 1.04% (95% CI: 0.22%–4.89%, n = 2 studies), 1.21% (95% CI: 0.75%–1.93%, n = 17 studies), 1.79% (95% CI: 1.36%–2.35%, n = 34 studies), 1.86% (95% CI: 1.41%–2.45%, n = 41 studies), 2.35% (95% CI: 0.01%–83.49%, n = 3 studies), respectively.

Conclusion

We provide robust point effect estimates of post‐ERCP cholangitis. Given the mortality and morbidity burden associated with ERCP‐related infectious outbreaks, this finding can inform public health policy regarding the health effects of hospital‐related infection on taking appropriate measures, especially in vulnerable populations.

1. Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) is a procedure that combines x‐ray and the use of upper gastrointestinal (GI) endoscopy for the diagnosis and treatment of pancreatic and biliary diseases. Technological refinements have established ERCP as a comparatively safe intervention. Initially conceived as a diagnostic modality in the late 1960s, its evolution into a therapeutic platform was catalyzed by the inaugural biliary sphincterotomy performed in 1974 Tokyo [1, 2, 3, 4, 5]. Contemporary practice predominantly employs ERCP for therapeutic indications, with more than 650 000 annual procedures in the USA [5, 6].

Procedure‐related complications occur in 4%–16% of cases, encompassing asymptomatic hyperamylasemia, cardiopulmonary depression, hypoxia, aspiration, intestinal perforation, bleeding, cholangitis, adverse medication reactions, sepsis, acute pancreatitis and death [7]. With the increase of ERCP indications, more attention is paid to identifying and preventing complications. Cholangitis represents the predominant infectious sequela.

Duodenoscope‐related outbreaks constitute significant iatrogenic threats, with documented potential 1‐month mortality rates of over 20% [8, 9]. Investigations confirm nosocomial transmission of multidrug‐resistant organisms (e.g., carbapenem‐resistant Enterobacteriaceae) via contaminated instrumentation [10, 11]. However, as ERCP is increasingly being utilized for different advanced techniques and being reported outbreak infections in several articles, less articles have been noticed about in‐depths analysis on infection complications with a lack of true point estimate on post‐ERCP infection. This knowledge gap persists due to: (1) heterogeneous diagnostic criteria across studies; (2) methodological disparities in surveillance; (3) inadequate representation of diverse populations.

Acute cholangitis portends substantial morbidity, mortality and healthcare utilization [12]. Standardized management protocols exist (ASGE, ESGE, Tokyo Guidelines), yet evidence characterizing ERCP‐attributable cholangitis remains limited [13, 14, 15]. Cholangitis are well‐known complication of ERCP. Clinical presentation includes fever, jaundice, and abdominal pain, and occasionally hypotension and altered mental status in severe cases [16].

This meta‐analysis addresses a critical evidence void through a comprehensive systematic review and quantitative synthesis of global epidemiological studies to determine post‐procedural cholangitis incidence and inform hospital related infection prevention and control frameworks.

2. Methods

2.1. Search Strategy

We conducted a systematic search by using both PubMed and EMBASE databases to identify epidemiologic studies that evaluated the incidence of cholangitis in post‐ERCP. We restricted our search to all‐language studies that were published up to February 6th, 2023.

2.2. Study Selection

We used the following search terms and keywords: (“ERCP” OR “Endoscopic retrograde cholangiopancreatography” OR “post‐ERCP”) AND (“cholangitis”) AND (“epidemiology” OR “epidemiological” OR “epidemiologic” OR “descriptive study” OR “observational study” OR “incidence”). We restricted our search to human studies. Synonyms of cholangitis and ERCP were included by using Medical Subheadings (MeSH) terms.

2.3. Data Extraction and Study Selection

Data extraction and accuracy assessment were done by authors from February 2023 to July 2023. Initial screening of abstracts was performed independently by four authors working in paired teams. Discrepancies in study inclusion were resolved through monthly consensus meetings chaired by the senior investigator with final arbitration authority. Full‐text assessment followed the same dual‐reviewer protocol. Extracted information was entered into an Excel database, which included titles, authors' names, publication year, continent, country, study period, sample size, numbers of post‐cholangitis, and age distribution. Definitive diagnosis of cholangitis followed the Tokyo Guidelines 2018 (TG18) criteria [15], requiring concurrent evidence of: (1) systemic inflammation (fever ≥ 38°C/chills or abnormal WBC/CRP); (2) cholestasis (jaundice or elevated liver enzymes > 1.5 × ULN); and (3) imaging‐confirmed biliary pathology. We excluded in vitro studies, book chapters, commentaries, letters to the editor, conference abstracts, toxicity studies, review articles, meta‐analyses and studies that were not written in English. To address duplicate records across databases, automated deduplication was performed using EndNote X9 identification function. For studies considered as duplicated cohorts, the publication with the most comprehensive temporal coverage and recent data was prioritized. The included articles are population who have performed ERCP in all levels' medical institutions. We focused on the incidence rate of cholangitis in post‐ERCP.

2.4. Statistical Analysis

All the incidence rates were calculated by using incidence number of cholangitis divided into study population. Overall and subgroup pooled effect estimates (continents, age, and publication year) as well as 95% confidence interval were evaluated by using inverse‐variance pooling and logit‐transformed proportions. Elderly group was considered population aged over 60 years old. Publication year subgroup analysis was separated as 10 years a group continually. We also tested heterogeneity for the reported effect estimates and provided the p‐values of the I ²‐based Cochran Q test and the I ² metric of inconsistency [17]. I ² > 50% are considered as substantial heterogeneity [17]. Sensitivity analyses were also conducted by adding back articles that accessed as “serious” under risk of bias assessment and rerun the model. Influence analysis was also conducted by omitting one study at a time to calculate remain overall pooled effect estimates. Forest plots were used to briefly display study information and proportion as well as influence analysis results for each study graphically. Statistical significance was assessed at the α = 0.05 level, unless otherwise reported. All statistical analyses were conducted in R version 4.0.1 using packages “meta,” “dmetar,” and “robvis.”

2.5. Risk of Bias Assessment

Risk of bias assessment for selected studies was conducted by using “Risk Of Bias In Non‐randomized Studies of Interventions (ROBINS‐I)” under seven domains (Bias due to confounding, Bias in selection of participants into the study, Bias in classification of interventions, Bias due to deviations from intended interventions, Bias due to missing data, Bias in measurement of outcomes, Bias in selection of the reported result) [18]. If any single domain is rated “medium,” “serious” or “critical” then the overall domain is rated similarly. The response options for an overall ROBINS‐I judgment are “low,” “moderate,” “serious,” “critical,” and “no information.” Studies with “serious” or “critical” overall risk ratings were excluded in primary analyses but underwent sensitivity analysis to assess their impact on pooled estimates.

3. Results

3.1. Characteristics of the Eligible Studies

Our study selection process is presented in Figure 1, which represents the PRISMA Flowchart. A total of 302 peer‐review articles were identified for our search in PubMed and EMBASE. Of these, after title and abstract screening, we identified a total of 122 articles that fulfilled our initial inclusion criteria, of which 285 were excluded because of the risk of bias assessment accessed as serious. As a result, 97 studies were included in the final meta‐analysis, comprising 17 studies from North America, 34 studies from Europe, 41 studies from Asia, 3 studies from Africa, and 2 studies from South America (Table 1). Table 1 summarizes detailed characteristics for the studies included in the final meta‐analysis.

FIGURE 1.

Flow chart of the study selection process.

TABLE 1.

Descriptive characteristics of included studies.

Continent	Country	Study	Study period	Total ERCP numbers	Numbers of post‐cholangitis	Study population	Mean age (SD) or range (years)
South America	Colombia	Peñaloza‐Ramírez et al. (2009) [19]	2006–2007	372	5	Patients conduct ERCPs at San Jose Hospital	52.23 ± 19.4
	Brazil	Pereira et al. (2020) [20]	1997–2013	2137	21	Patients with known or suspected bile duct stones conduct ERCPs	57 (0.8–104)
North America	USA and Canada	Freeman et al. (1996) [21]	1992–1994	2347	24	Patients conduct ERCPs at 16 institutions	All ages
North America	USA	Nebel et al. (1975) [22]	1974	404	25	1974 A/S/G/E survey	All ages
North America	USA	Bilbao et al. (1976) [23]	1974	8681	72	United States owners of side‐viewing duodenoscopes survey	All ages
North America	USA	Tham et al. (1997) [24]	1995	190	3	Patients conduct ERCP as inpatients	All ages
North America	USA	Cheng et al. (2005) [25]	1994–2003	329	1	Patients in Whitcomb Riley Hospital for Children	17 years of age or younger
North America	USA	Gluck et al. (2008) [26]	1984–2005	317	3	Patients underwent ERCP at Virginia Mason Medical Center, Seattle	47 (15–86)
North America	USA	Bangarulingam et al. (2009) [27]	2005	1149	8	All Mayo Clinic patients who underwent ERCP	All ages
North America	USA	Adler et al. (2010) [28]	2004–2008	801	2	Patients undergoing ERCP with the primary purpose of obtaining access to the biliary tree at the University of Texas, Health Science Center and the University of Utah, Health Science Center	46 (18–96)
North America	USA	Alkhatib et al. (2011) [29]	2000–2009	185	2	Patients with a diagnosis of PSC who underwent ERCP at academic institutions	All ages
North America	USA	Sethi et al. (2011) [30]	2001–2007	4214	11	Patients conduct ERCP at University of Colorado Hospital	All ages
North America	USA	Coelho‐Prabhu et al. (2013) [31]	1997–2006	1072	16	All adult residents of Olmsted County, Minnesota, who underwent ERCP	57.6 ± 19.8
North America	USA	Adler et al. (2016) [32]	2003–2014	538	15	Patients with cirrhosis who underwent a therapeutic ERCP in all centers	53.2 ± 14.4
North America	USA	Al‐Mansour et al. (2018) [33]	2003–2016	2392	11	Patients who underwent ERCP procedures at Ohio State University and Southern Illinois University	53.4 (7–102)
North America	USA	Jang et al. (2018) [34]	2008–2016	645	27	Patients who underwent ERCP with SEMS placement for the management of a malignant bile duct stricture	68.6 ± 12.4
North America	USA	Cassani et al. (2019) [35]	1997–2014	485	11	Patients with perihilar cholangiocarcinoma who underwent ERCPs	63 (21–89)
North America	USA	Kwak et al. (2020) [36]	2007–2017	1079	21	Patients 18 years of age or older that underwent ERCP at the University Hospital of Brooklyn (UHB) at SUNY Downstate Medical Center and Kings County Hospital Center (KCHC)	60.7 (18–94)
North America	Canada	Mohamed et al. (2021) [37]	2018–2020	637	8	Patients undergoing ERCP for biliary indications	All ages
Europe	UK	Williams et al. (2007) [38]	2004	4561	48	Patients aged over 17 years conduct ERCP at 66 centers	65.0 ± 16.7
Europe	UK	Chatterjee et al. (2011) [39]	2009	481	4	Northern Regional Endoscopy Group (NREG) study	69
Europe	Switzerland	Sulz et al. (2012) [40]	2005–2007	471	2	All patients undergoing EST during ERCP at department of Internal Medicine, Division of Gastroenterology and Hepatology	All ages
Europe	Sweden	Lübbe et al. (2015) [41]	2007–2012	35 944	967	Swedish Registry for Gallstone Surgery and ERCP (GallRiks)	All ages
Europe	Spain	García‐Cano Lizcano et al. (2004) [42]	1997–2002	507	6	Patients conduct ERCP at Hospital Virgen de la Luz	All ages
Europe	Spain	Balderramo et al. (2011) [43]	2003–2010	243	8	All ERCPs performed in LT recipients in our institution	All ages
Europe	Spain	Gómez‐Oliva et al. (2012) [44]	2009	199	2	Nonrandomized database of consecutive patients with malignant biliary extrahepatic obstruction and contraindication to surgical resection in 32 Spanish hospitals	75 ± 12
Europe	Spain	Leal et al. (2019) [45]	2002–2015	283	5	Patients undergoing ERCP at selected hospitals	68.43 ± 14.8
Europe	Romania, Italy, Croatia, Romania, Croatia, Serbia	Voiosu et al. (2020) [46]	2016–2018	1843	45	Sixhigh and low volume centers across Europe	66.8 ± 14.6
Europe	Romania	Voiosu et al. (2016) [47]	2014–2015	534	5	Patients performed ERCP at Colentina Clinical Hospital	64 ± 15.3
Europe	Portugal	Sousa et al. (2018) [48]	2014–2016	113	10	Patients with more than 75 years underwent ERCP	82 ± 7
Europe	Portugal	Damiao et al. (2021) [49]	2016–2019	1046	7	Consecutive ERCP procedures registered in a database at our endoscopy unit	75.7 (18–100)
Europe	Portugal	Rodrigues‐Pinto et al. (2021) [50]	2012–2017	2002	97	Consecutive patients who underwent ERCPs	68 (56–79)
Europe	Poland	Kostrzewska et al. (2011) [51]	2001–2004	734	7	ERCP procedures performed in patients by 3 endoscopists in the Endoscopy Unit in the Department of Gastroenterology and Internal Medicine	62.6 (19–99)
Europe	Norway	Glomsaker et al. (2013) [52]	2007–2009	2808	100	All patients aged 18 years or more scheduled for ERCP were included in the study	All ages
Europe	Netherlands	Schreurs et al. (2002) [53]	1985–1995	552	8	Consecutive patients with symptomatic cholelithiasis were treated at St. Elisabeth Hospital	All ages
Europe	Netherlands	Klaske et al. (2014) [54]	1990–2012	192	3	Patients were referred for BDI treatment and included in a prospective database	51.8 ± 15.7
Europe	Italy	Coppola et al. (1997) [55]	1988–1995	546	7	Consecutive patients underwent endoscopic retrograde cholangiography (ERCP) for biliary stone	63.7
Europe	Italy	Loperfido et al. (1998) [56]	1992–1994	2769	24	Consecutive patients undergoing ERCP at nine centers in the Triveneto region of Italy	66 (7–93)
Europe	Italy	Bove et al. (2015) [57]	1982–2012	713	5	Patients with Billroth II reconstruction who underwent ERCP	All ages
Europe	Italy	Galeazzi et al. (2018) [58]	2013–2015	465	23	Patients aged ≥ 65 undergoing ERCP at San Gerardo Hospital ASST Monza	77.6 ± 7.64
Europe	Italy	Donato et al. (2021) [59]	2016–2018	766	8	Nineteen Italian centers performing ERCP	All ages
Europe	Italy	Fugazza et al. (2022) [60]	2016–2019	369	14	All physicians who had per formed at least one DSOC were contacted and 18 centers in Italy accepted to be part of this prospective evaluation	66.7 ± 14.7
Europe	Greece	Christoforidis et al. (2008) [61]	1998–2006	272	6	Patients underwent ERCP for suspected choledocholithiasis	79 (75–89)
Europe	Greece	Papaefthymiou et al. (2022) [62]	2015–2020	1082	43	Patients hospitalized in the Department of Gastroenterology of the General University Hospital of Larissa	72.7 ± 15.82
Europe	Germany	Peiseler et al. (2018) [63]	2009–2017	663	15	Patients with PSC managed ERCP at University Medical Centre Hamburg Eppendorf	No information
Europe	France	Barthet et al. (2002) [64]	1996–2000	1159	20	Patients conduct ERCP at Department of Gastroenterology, Hospital Nord	All ages
Europe	France	Vitte et al. (2007) [65]	1999–2000	2708	51	All patients undergoing an upper gastroenterological endoscopic procedure with planned catheterization of the biliary or pancreatic ducts were included	70.1 ± 17
Europe	Finland	Salminen et al. (2008) [66]	1997–2005	2555	2	Patients underwent ERCP at Turku University Central Hospital	All ages
Europe	Finland	Ismail et al. (2012) [67]	2007–2009	441	6	Patients conduct ERCPs with PSC	All ages
Europe	Finland	Siiki et al. (2012) [68]	2002–2009	1207	17	EECPs performed in the endoscopy unit of KantaHäme Central Hospital (KHCH) in Hämeenlinna	All ages
Europe	Denmark	Ambrus et al. (2015) [69]	2003–2012	292	12	Ltx patients conduct ERCP at University Hospital of Copenhagen	All ages
Europe	Denmark	Tan et al. (2018) [70]	2009–2016	166	1	All patients undergoing ERCP at OUH	71 ± 9
Europe	Austria	Kapral et al. (2012) [71]	2006–2011	13 511	185	A Nationwide voluntary ERCP benchmarking project that was initiated by the Austrian Society of Gastroenterology and Hepatology	All ages
Asia	Turkey	Koklu et al. (2005) [72]	2002–2003	202	2	Patients had naive papilla in Türkiye Yüksek Ihtisas Hospital	49.83 ± 13.55
Asia	Turkey	Gozel et al. (2016) [73]	2009–2012	339	11	Patients who underwent ERCP at Akdeniz University Medical School Department of Gastroenterology	No information
Asia	Turkey	Yıldırım et al. (2017) [74]	2010–2014	1044	14	Patients who underwent ERCP for the first time and are over the age of 65 at Gaziantep University Hospital	≥ 65
Asia	Turkey	Cankurtaran et al. (2021) [75]	2019–2020	341	2	Patients underwent ERCP at Ankara City Hospital	58.86 ± 18.41
Asia	Thailand	Pungpapong et al. (2005) [76]	2000–2002	584	35	Patients conduct ERCP at Chulalongkorn University Hospital	16–97
Asia	Thailand	Laohavichitra et al. (2007) [77]	2003–2004	416	2	Patients conduct ERCP in Siriraj GI Endoscopy Center, Faculty of Medicine, Siriraj hospital, Mahidol University, Bangkok, Thailand	60.6 (14–97)
Asia	Singapore	Chong et al. (2005) [78]	1999–2002	144	2	Patients aged 80 years and over and underwent ERCP at the Gastroenterology Unit, Tan Tock Seng Hospital	84.64 ± 3.90
Asia	Singapore	Ong et al. (2005) [79]	2001–2003	336	9	Patients conduct ERCP in National University Hospital, Singapore	No information
Asia	Nepal	Gurung et al. (2014) [80]	2011–2013	516	6	All the patients done since August 2011 to August 2013 were retrieved	50.57 ± 17.78
Asia	Korea and Japan	Park et al. (2013) [81]	2004–2010	946	6	Consecutive patients who underwent attempted removal of CBD stones C10 mm in size using EPLBD	72.7 ± 11.0
Asia	Korea	Jang et al. (2010) [82]	1994–2008	245	1	Pediatric patients who underwent ERCPs in the Asan Medical Center	8.0 ± 4.2
Asia	Korea	Kim et al. (2016) [83]	2009–2014	536	4	Patients with naive papilla who had undergone ERCP for treatment of CBD stone Yeungnam University Hospital	68.6 ± 15.2
Asia	Korea	Lee et al. (2020) [84]	2015	955	11	Patients with naive papilla who underwent ERCP at six centers	71 (19–101)
Asia	Korea	Choi et al. (2021) [85]	2011–2018	483	17	Patients who had undergone endoscopic CBD stone removal by ERCP	61.4 ± 17.2
Asia	Japan	Sugiyama et al. (2000) [86]	1977–1997	403	6	Patients underwent ERCP for choledocholithiasis	70–96
Asia	Japan	Tsujino et al. (2005) [87]	2002–2004	202	2	Patients in 4 endoscopic centers in Tokyo	65 (22–92)
Asia	Japan	Nakahara et al. (2009) [88]	1998–2008	808	8	Patients who underwent ERCP at Sendai City Medical Center	74.1 (26–101)
Asia	Japan	Nishikawa et al. (2014) [89]	2004–2012	743	16	Consecutive patients (age ≥ 20 years) with biliary stone diseases, including common bile duct stones, Mirizzi syndrome type II stones, and intrahepatic stones, were retrospectively analyzed	70.0 ± 10.9 (27–95)
Asia	Japan	Tohda et al. (2016) [90]	2008–2016	207	3	Patients performed emergency ERCPs at Fukui Kosei Hospital	58–92
Asia	Japan	Saito et al. (2019) [91]	2012–2018	695	13	Patients who had native papilla underwent ERCP	75–104
Asia	Japan	Saito et al. (2021) [92]	2012–2020	1491	23	Patients with native major duodenal papilla diagnosed with choledocholithiasis	73.7 ± 10.2
Asia	Japan	Tanisaka et al. (2022) [93]	2011–2019	1318	19	Patients who underwent ERCP related procedures using short SBE	73 (66–79)
Asia	Iran	Hormati et al. (2019) [94]	2014–2018	1023	9	Patients conduct ERCP in Shahid Beheshti Hospital	47.2 ± 6.7
Asia	India	Jagtap et al. (2019) [95]	2012–2016	261	6	Patients with cirrhosis who underwent ERCP Asian Institute of Gastroenterology, Hyderabad, India	53.49 ± 12.57
Asia	China	Lo et al. (1997) [96]	1987–1994	706	14	Patients conduct ERCP at Queen Mary Hospital	19–99
Asia	China	Hu et al. (2009) [97]	2004–2007	303	3	Patients underwent ERCP at Mackay Memorial Hospital	≥ 65
Asia	China	Liao et al. (2009) [98]	2002–2007	812	34	Patients who underwent therapeutic ERCP during live demonstrations performed in 14 selected major endoscopy centers in the mainland of China	All ages
Asia	China	Wang et al. (2009) [99]	2006–2007	2691	38	Consecutive ERCP procedures were studied at 14 centers in China	57.88 ± 17.43
Asia	China	Wang et al. (2016) [100]	2010–2014	121	3	Patients conduct ERCP Beijing Military General Hospital	83.65 ± 3.45 (80–95)
Asia	China	Wan et al. (2018) [101]	2010–2016	252	30	Patients diagnosed with unresectable cholangiocarcinoma and who had undergone ERCP at Shanghai General Hospital	72.06 ± 9.83
Asia	China	Zhang et al. (2016) [102]	2008–2015	532	5	Patients with intact papilla who were established as candidates for therapeutic ERCP at Shanghai First People's Hospital	69.2 ± 11.7 (32–91)
Asia	China	Chen et al. (2017) [103]	2009–2015	1500	22	Patients conduct ERCP in Ruijin Hospital	57.98 ± 14.89
Asia	China	Du et al. (2017) [104]	2012–2015	1734	70	Patients conduct ERCP Chinese PLA General Hospital	61 (10–98)
Asia	China	Li et al. (2019) [105]	2002–2016	391	8	Consecutive patients with a history of Billroth II gastrectomy who underwent ERCP	66.6 ± 12.3
Asia	China	Tabak et al. (2020) [106]	2016–2018	614	4	Patients with a native papilla	68 (54–79)
Asia	China	Xia et al. (2020) [107]	2002–2018	502	108	Acute PEC in patients with inoperable MHBS	62.4 ± 12.2
Asia	China	Xu et al. (2020) [108]	2016–2019	327	12	Patients with native papilla were invited to participate into the study	No information
Asia	China	Li et al. (2022) [109]	2003–2019	567	25	All patients with and without cirrhosis who underwent therapeutic ERCP	All ages
Asia	China	Liu et al. (2022) [110]	2014–2021	345	24	Patients with endoscopic biliary drainage in our Zhongshan Hospital of Traditional Chinese Medicine	67.8 ± 11.2
Asia	China	Zhang et al. (2022) [111]	2002–2021	156	1	Patients aged ≥ 85 years with biliary stones who underwent their first ERCP at Chinese PLA General Hospital	87 (86–89)
Asia	China	Zhou et al. (2022) [112]	2011–2020	20 652	227	All procedures performed ERCP during this period were included in this study	All ages
Africa	South Africa	Lubbe et al. (2022) [113]	2015–2020	153	33	Consecutive patients in whom an index drainage procedure was performed for malignant hilar obstruction with palliative intent at the specialized referral centers of Groote Schuur and Tygerberg Hospitals	61.8 ± 12.5
Africa	Libya	Tumi et al. (2015) [114]	2005–2010	759	4	Consecutive ERCP procedures were performed in 704 patients at the Endoscopy Unit of Central Hospital, Tripoli, Libya	56.8 ± 18.7
Africa	Egypt	Omar et al. (2015) [115]	2010–2015	908	8	Consecutive patients who underwent an ERCP	41.4 ± 9.5

3.2. Risk of Bias Assessment

The summary of the risk of bias assessment is shown in summary plot (Figure 2). In two out of seven domains (bias in classification of interventions, bias due to deviations from intended interventions), the risk of bias assessment was found to be only low. While in two out of seven domains (bias due to missing data, bias in selection of the reported result), the risk of bias assessment was found to be low or moderate. In the remaining three domains (bias due to confounding, bias due to selection of participants and bias in measurement of outcomes), we found a variable proportion of articles having serious risk of bias. Three, 22, and one articles were classified as serious in the bias due to confounding domain, bias due to selection of participants domain and bias in measurement of outcomes domain, respectively. Twenty‐five articles (20.5%) were assessed as serious in overall risk of bias assessment and were excluded in the main meta‐analysis. Risk of bias assessment for detailed individual studies is shown in the traffic plot (Figure S1a–c).

FIGURE 2.

Summary of risk of bias assessment.

3.3. Results of the Meta‐Analysis

Table 2 and Figure 3a,b present the pooled effect estimates as well as heterogeneity for ERCP post cholangitis. The pooled overall rate was 1.70% (95% CI: 1.41%–2.04%, I ² = 94.4). The pooled rate for South America, North America, Europe, Asia, and Africa were 1.04% (95% CI: 0.22%–4.89%, n = 2 studies, I ² = 0), 1.21% (95% CI: 0.75%–1.93%, n = 17 studies, I ² = 91.2), 1.79% (95% CI: 1.36%–2.35%, n = 34 studies, I ² = 90.3), 1.86% (95% CI: 1.41%–2.45%, n = 41 studies, I ² = 95.6), 2.35% (95% CI: 0.01%–83.49%, n = 3 studies, I ² = 98.1), respectively. The pooled rates for age subgroup were 1.66% (95% CI: 1.36%–2.02%, n = 86 studies, I ² = 94.9) for all ages and 2.15% (95% CI: 1.31%–3.51%, n = 11 studies, I ² = 75.9) for elderly. The pooled rates for publication year subgroup were 2.28% (95% CI: 0%–99.99%, n = 2 studies, I ² = 98.7) for 1970s, 1.20% (95% CI: 0.77%–1.85%, n = 5 studies, I ² = 41.7) for 1990s, 1.36% (95% CI: 0.95%–1.95%, n = 22 studies, I ² = 84.6) for 2000s, 1.66% (95% CI: 1.28%–2.14%, n = 45 studies, I ² = 89.9) for 2010s, and 2.35% (95% CI: 1.49%–3.69%, n = 23 studies, I ² = 97.6). Figures 4 and 5 were displayed the global distribution of incidence number as well as incidence rate of post‐ERCP cholangitis visually by using world map.

TABLE 2.

Pooled incidence rate of ERCP post‐cholangitis.

	Studies, n	Incidence % (95% CI)	I 2, %
Full meta‐estimate	97	1.70 (1.41, 2.04)	94.4
Continent
South America	2	1.04 (0.22, 4.89)	0
North America	17	1.21 (0.75, 1.93)	91.2
Europe	34	1.79 (1.36, 2.35)	90.3
Asia	41	1.86 (1.41, 2.45)	95.6
Africa	3	2.35 (0.01, 83.49)	98.1
Age
Elderly	11	2.15 (1.31, 3.51)	75.9
All ages	86	1.66 (1.36, 2.02)	94.9
Publication year
1970s	2	2.28 (0.00, 99.99)	98.7
1990s	5	1.20 (0.77, 1.85)	41.7
2000s	22	1.36 (0.95, 1.95)	84.6
2010s	45	1.66 (1.28, 2.14)	89.9
2020s	23	2.35 (1.49, 3.69)	97.6

FIGURE 3.

Forest plot of study‐specific incidence rate of post‐ERCP cholangitis. *The meta‐estimate and weights in the forest plot are estimated from random effects meta‐analyses.

FIGURE 4.

Worldwide distribution of post‐ERCP cholangitis numbers for countries.

FIGURE 5.

Worldwide distribution of incidence rate of post‐ERCP cholangitis. *For countries with more than one studies, results were pooled by meta‐analysis.

3.4. Sensitive Analysis

Table 3 summarized the detailed pooled effect estimated by adding back articles that accessed as “serious” under risk of bias assessment. The pooled overall rate for sensitive analysis was 1.99% (95% CI: 1.64%–2.40%, n = 122 studies, I ² = 95.4). The pooled rate for South America, North America, Europe, Asia, and Africa were 1.04% (95% CI: 0.22%–4.89%, n = 2 studies, I ² = 0), 1.38% (95% CI: 0.83%–2.30%, n = 20 studies, I ² = 95.5), 2.17% (95% CI: 1.60%–2.93%, n = 47 studies, I ² = 93.8), 2.10% (95% CI: 1.57%–2.81%, n = 50 studies, I ² = 96.1), 2.35% (95% CI: 0.01%–83.49%, n = 3 studies, I ² = 98.1), respectively. The pooled rates for publication year subgroup were 2.28% (95% CI: 0%–99.99%, n = 2 studies, I ² = 98.7) for 1970s, 2.18% (95% CI: 0.78%–5.96%, n = 8 studies, I ² = 95.1) for 1990s, 1.47% (95% CI: 1.09%–2.00%, n = 26 studies, I ² = 81.8) for 2000s, 1.96% (95% CI: 1.48%–2.59%, n = 56 studies, I ² = 93.3) for 2010s, and 2.64% (95% CI: 1.69%–4.10%, n = 30 studies, I ² = 97.8). The 0.29% absolute incidence increase in sensitivity analysis primarily originated from historical studies and underpowered regional cohorts. Crucially, exclusion criteria removed studies contributing disproportionately to variance, reducing I ² from 95.4% to 94.4% in the primary analysis.

TABLE 3.

Sensitive analysis of ERCP post‐cholangitis.

	Studies, n	Incidence (95% CI)	I 2, %
Full meta‐estimate	122	1.99 (1.64, 2.40)	95.4
Continent
South America	2	1.04 (0.22, 4.89)	0
North America	20	1.38 (0.83, 2.30)	95.5
Europe	47	2.17 (1.60, 2.93)	93.8
Asia	50	2.10 (1.57, 2.81)	96.1
Africa	3	2.35 (0.01, 83.49)	98.1
Publication year
1970s	2	2.28 (0.00, 99.99)	98.7
1990s	8	2.18 (0.78, 5.96)	95.1
2000s	26	1.47 (1.09, 2.00)	81.8
2010s	56	1.96 (1.48, 2.59)	93.3
2020s	30	2.64 (1.69, 4.10)	97.8

Figure S2a,b displayed detailed forest plot of the recalculated overall effect with one study omitted each time. The figure visually summarizes the overall effect estimate changes when different studies are removed. No obvious outlier is detected in the study.

4. Discussion

In this systematic review, we identified 97 studies that examined the prevalence of post‐ERCP cholangitis in all levels of hospitals all over the world. Our study examined 158 434 ERCP operations worldwide with which 2949 cases of post‐ERCP cholangitis were observed in total. Our study reported the incidence rate of post‐ERCP cholangitis ranging from 0.08% to 21.57%, with the pooled overall rate being 1.70% (95% CI: 1.41%–2.04%). Sensitivity analysis incorporating articles accessed as having serious risk of bias yielded nearly identical results.

One previous systematic review by Deb et al. [11] integrated studies published from 2011 to 2020, reporting a composite post‐ERCP infection rate of 0.8% (3452 out of 433 414 procedures), encompassing all infectious complications where cholangitis and sepsis comprised over 77% of cases. However, cholangitis was not analyzed as a distinct weighted outcome, as the study simply divided the number of incidences by the total number without weighting. Three meta‐analyses were performed. Forbes et al. [116] reported a pooled cholangitis incidence of 2.5% (95% CI: 1.9%–3.3%) across 121 619 procedures predominantly from high‐income regions, noting high heterogeneity across the groups, which is consistent with our study. The remaining two meta‐analyses addressed ERCP‐related complications in restricted populations: one in liver transplant recipients reporting a cholangitis incidence of 0.8% (95% CI: 0.4%–2.0%) [117] and another in choledocholithiasis patients demonstrating a pooled cholangitis risk of 3.09% in asymptomatic versus 2.36% in symptomatic subgroups [118]. Compared to previous reviews focusing on multiple complications, our analysis of 158 434 ERCP cases provides granular geographic stratification using a larger global population, thereby overcoming sampling gaps.

Furthermore, whereas previous analysis combined various post‐ERCP infection endpoints (cholangitis, cholecystitis, Carbapenem‐Resistant Enterobacteriaceae [CRE] infection, etc.), our study specifically focused on the most common infection endpoint, cholangitis, and employed a weighting procedure to generate more robust results.

Geographical analysis revealed the lowest point estimate was observed in South America (1.04%, 95% CI: 0.22%–4.89%) with only two published articles pooled and the highest articles were reported in Africa (2.35%, 95% CI: 0.01%–83.49%) with three published articles pooled. The pooled incidence rate in Europe (1.79%, 95% CI: 1.36%–2.35%) was slightly lower than the results in North America (1.21%, 95% CI: 0.75%–1.93%) or in Asia (1.86%, 95% CI: 1.41%–2.45%). The remarkably wide confidence intervals observed for South America (0.22%–4.89%) and Africa (0.01%–83.49%) reflect substantial methodological heterogeneity beyond true epidemiological variation. This variability may stem from inconsistent application of cholangitis diagnostic criteria—particularly variable adherence to Tokyo Guidelines across settings—combined with divergent surveillance protocols among healthcare systems. These limitations are further compounded by limited sample sizes (n = 2 and n = 3 studies respectively), which amplify statistical uncertainty, and potential systematic underreporting in resource‐constrained regions where post‐procedural monitoring infrastructure may be suboptimal. The pooled results in elderly population were higher than in all age population, which is coherent with a scoping literature review that published in 2019 that identified age as a risk factor for post‐ERCP cholangitis [119]. The significantly elevated incidence of post‐ERCP cholangitis in elderly patients warrants specific clinical adaptations for this vulnerable population. Current evidence supports implementing extended 24‐h hemodynamic monitoring with structured fever response protocols, mandating blood cultures and CRP testing within 6 h of temperature exceeding 38°C to address accelerated sepsis progression in geriatric patients [16]. Antibiotic prophylaxis should be reserved exclusively for high‐risk scenarios including incomplete biliary drainage, liver transplant recipients, and post‐cholangioscopy biopsy cases, utilizing second‐generation cephalosporins targeting biliary flora [16, 120, 121, 122, 123]. Procedural technique optimization remains paramount, particularly avoiding mechanical lithotripsy in favor of endoscopic papillary large‐balloon dilation for stone extraction to minimize retained contaminants to prevent iatrogenic inoculation [16, 123, 124]. Furthermore, we observe that the point estimates show an increasing trend as the year of publication changes. This indicates that despite the continuous innovation of technology and the continuous optimization of endoscopes and disinfestation procedures, there are many other factors that affect infection which need further analysis. Our sensitivity analysis incorporating studies with serious risk of bias yielded a pooled incidence of 1.99% (95% CI: 1.64%–2.40%)—representing a 17.1% relative increase from the primary estimate of 1.70%. This divergence primarily originates from two variance hotspots: African cohorts demonstrated extreme dispersion (2.35% [0.01%–83.49%], I ² = 98.1) reflecting surveillance limitations in undersampled regions (n = 3), while 1990s‐era studies showed 81.7% inflation (2.18% vs. primary 1.20%) indicative of historical diagnostic imprecision. Crucially, the substantial confidence interval overlaps between analyses (primary 1.41%–2.04% vs. sensitivity 1.64%–2.40%) confirms core epidemiological coherence, affirming that methodological rigor enhances precision without altering fundamental risk characterization.

Cholangitis is the most common infection complication with which risk factors include old age, previous ERCP history, hilar biliary obstruction and disinfection procedure [119]. Chen et al. [125] identified hilar obstruction as an independent predictor (OR 2.586; 95% CI: 2.066–2.743), attributing this to tumor‐induced isolation of intrahepatic ducts, which impedes complete drainage and promotes bacterial colonization. Cotton et al. [122] demonstrated that liver transplant recipients face significantly elevated risks (1.2% vs. 0.25% in non‐transplant patients) due to biliary‐enteric anastomoses facilitating bacterial reflux and immunosuppression impairing infection clearance. Infection prevention in endoscopy has been a focus of research for many years. Antimicrobial prophylaxis, adherence to adequate endoscope disinfection procedures, improved screening techniques for detection of endoscope contamination, and the use of disposable endoscopes are a few strategies that have been widely proposed and studied for the prevention of infection after GI endoscopy [11].

There are two common mechanisms for post‐ERCP cholangitis. One mechanism of infection involves colonization by bacteria or endotoxins that penetrate the bile and cross the blood barrier in the setting of biliary stasis or during a prolonged ERCP operation [15]. Another mechanism is damaging the epithelium during contrast injection or other conduits that allow bacterial entry. Moreover, previously placed stents may also become obstructed (due to stone fragments, bacterial biofilm, sludge, tumor or tissue growth) and block the lumen of the stent, resulting in delayed infection [16].

Our study observed a large degree of heterogeneity for all cholangitis outcomes pairs across enrolled studies, indicating a significant variation among results that could not be expected by chance alone. Although such heterogeneity does not impact our determination of consistency in causal inference, it is still essential to explore why the results are so disparate with each other [126]. Potential high heterogeneity that we observed may be associated with location, levels of hospital, different use of disinfectants as well as different operators, and so forth. The potential possible detailed stated as follows:

Duodenoscopes that are used by ERCP operation are precision instruments that invade the human body. Due to their complex structure and unique shape, a high rate of persistent bacterial contamination exists even after automated reprocessing and disinfection, posing a certain risk of cross infection for patients. Inadequate cleaning of flexible endoscopes before disinfection can spread deadly pathogens rank as top 1 in 2016 top 10 Health Technology Hazards released by the American Emergency Care Research Institute (ERCI) [127]. Due to the unique structure and material of soft endoscopes, it is easy to form biofilms within the lumen of the soft endoscope. The residual organic matter and water in the lumen can further promote the growth of biofilms, which have strong resistance to the external environment and often accumulate gradually during repeated use, treatment, and reuse. Bacteria are coated in the biofilm, and when it reaches a certain thickness, it affects disinfection and sterilization. Achieving disinfection and sterilization through the use of disinfectants and sterilization practices is essential for ensuring that medical and surgical instruments do not transmit infectious pathogens to patients [128]. Moreover, recent outbreaks of duodenoscope‐associated multidrug‐resistant organism infections have increased awareness and concern about the pitfalls in repeat high‐level disinfection protocols [129]. All kinds of disinfectants have different degrees of restrictions, and the rational use of disinfectants needs to be further studied in the future.

Global heterogeneity exists in endoscopist competency requirements. In the United States, ERCP training requires 3 years of internal medicine, 3 years of general gastroenterology, and 1 year of advanced endoscopy training to develop essential cognitive competencies including therapeutic decision‐making and core technical skills [5, 6, 130]. However, real‐world outcomes remain compromised by systemic challenges. Most critically, 89% of practicing endoscopists perform fewer than 25 ERCPs annually, a volume threshold repeatedly linked to higher procedural failure rates and hospitalization risks [130]. This low‐volume reality is exacerbated by inconsistent institutional oversight, fewer than half of U.S. healthcare facilities enforce re‐credentialing via minimum volumes or cannulation rates [5]. When coupled with inherent variations in operator learning curves and individual differences in non‐technical skills like situational awareness, these systemic gaps create fertile ground for infection rate disparities across practitioners.

We provide robust point effect estimates of post‐ERCP cholangitis. Given the mortality and morbidity burden associated with ERCP‐related infectious outbreaks, this finding can inform public health policy regarding the health effects of hospital‐related infection on taking appropriate measures, especially in vulnerable populations.

Funding

This work was supported by the Shanghai Pujiang Program (22PJD041), the Hospital Management Research Fund of Shanghai Hospital Association (X2023170), and the Three‐Year Initiative Plan (2023–2025) for Strengthening Public Health System Construction in Shanghai, Key Disciplines (GWVI‐11.1‐04).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Traffic light plot of risk of bias assessment for each study.

Figure S2: Forest plot of influence analysis.

Data Availability Statement

Research data are not shared.