Impact of body mass index on the outcomes of endoscopic retrograde cholangiopancreatography: a systematic review and network meta-analysis
Article information
Abstract
Background/Aims:
Obesity is an increasing health concern worldwide, and an elevated body mass index (BMI) may influence the outcomes of endoscopic retrograde cholangiopancreatography (ERCP). This systematic review and network meta-analysis aimed to investigate the association between BMI and ERCP outcomes.
Methods:
A systematic search was performed using PubMed, Scopus, and Web of Science to identify relevant studies that reported clinical outcomes of ERCP in different BMI groups. Patients were categorized into five BMI-based following groups: underweight (BMI<18.5 kg/m²), normal weight (18.5 kg/m²≤BMI<25 kg/m²), overweight (25 kg/m²≤BMI<30 kg/m²), obesity (30 kg/m²≤BMI<40 kg/m²), and morbid obesity (40 kg/m²≤BMI).
Results:
Among 3,001 unique citations, seven were included in the study. Meta-analysis revealed that obesity was not associated with an increased risk of post-ERCP pancreatitis (PEP) (odds ratio, 1.33; 95% confidence interval, 0.62–2.87). In the network analysis, none of the five BMI groups had a significantly higher risk of PEP than that of the other groups (p>0.05). Moreover, the BMI groups had similar rates of difficult cannulation, successful cannulation, complete stone extraction, and procedural success.
Conclusions:
ERCP success and adverse events were similar among different BMI groups. Neither an elevated nor a low BMI was associated with an increased risk of PEP. Further large-scale prospective studies are required to validate these findings.
INTRODUCTION
Obesity is a growing worldwide health issue; according to the World Health Organization, approximately 60% of adults were overweight or had obesity in 2022, and this rate is on the rise.1 Obesity is associated with a vast array of comorbidities including cardiovascular diseases, diabetes mellitus, and other gastrointestinal disorders.2 Among endoscopic procedures, obesity presents certain technical challenges, such as the risk of adverse events.3
Endoscopic retrograde cholangiopancreatography (ERCP) is an endoscopic procedure used to manage pancreatobiliary disorders. While technical improvements have significantly improved its safety profile, ERCP remains associated with major adverse events, such as post-ERCP pancreatitis (PEP), perforation, and bleeding. Patient-related factors such as age, comorbidities, and anatomical differences have been reported to influence ERCP.4 However, the specific impact of obesity on the technical success and adverse event rates of ERCP remains controversial.
Obesity can adversely affect ERCP outcomes in several ways. Visceral fat and abdominal girth may make patient positioning and scope manipulation more challenging, leading to increased procedure duration and technical difficulties. Patients with obesity also require higher doses of sedation or anesthesia, which increase the risk of cardiopulmonary complications.5 The altered anatomy and higher incidence of comorbidities, such as chronic liver disease and metabolic syndrome, can also contribute to an increased risk of infection and other complications.6 It is necessary to determine the effect of obesity on ERCP outcomes to optimize patient management, planning, and risk stratification.
As the obesity epidemic sweeps worldwide, endoscopists are encountering an increasing number of patients with obesity who require ERCP. This study investigated the effects of obesity and body mass index (BMI) on ERCP outcomes. This could help guide clinicians in preprocedural planning and prevention strategies by clarifying whether obesity plays an independent role in ERCP-related complications and procedural failure rates.
METHODS
Study protocol
This study was performed in accordance with Preferred Reporting Items for Systematic reviews and Meta-Analyses for Network Meta-Analyses (NMAs) guideline from study planning to the completion of the manuscript. The protocol is registered in the PROSPERO database (registration ID: CRD420251031419).
Data sources and search strategy
A systematic search was performed to identify all studies on the association between obesity and ERCP in PubMed, Scopus, and Web of Science databases up to April 2025. The main keywords used in the literature search were ‘obesity and endoscopic retrograde cholangiopancreatography’. Supplementary Table 1 presents detailed search strategies for each database. Additionally, manual screening of the reference lists of eligible studies was performed. No restrictions on language or date were imposed.
Study selection and eligibility criteria
The citations obtained through database search were combined, and duplicates were removed using EndNote X20 (Clarivate). Two authors (E.A. and S.K.) independently screened the citations by title/abstract and full-text to exclude those irrelevant to the study. All eligible studies included in our analysis met the PECOS criteria as follows: (1) Population: patients aged 18 years or older undergoing ERCP; (2) Exposure and Comparison: five BMI-based groups—underweight (BMI<18.5 kg/m²), normal weight (18.5 kg/m²≤BMI<25 kg/m²), overweight (25 kg/m²≤BMI<30 kg/m²), obesity (30 kg/m²≤BMI<40 kg/m²), and morbid obesity (BMI≥40 kg/m²); (3) Outcome: biliary cannulation success and difficulty, procedural success, and post-ERCP adverse events; and (4) Study design: prospective or retrospective studies. Studies that met any of the following criteria were excluded: studies without raw data, non-English articles, and articles for which the full-text was unavailable.
Data extraction
Two authors (E.A. and S.K.) independently collected the data using a pre-specified Excel ver. 2021 (Microsoft Corporation) spreadsheet. The following items were extracted from all eligible studies: first author, publication year, country of origin, study design and timeframe, inclusion and exclusion criteria, patient demographics (age and sex), investigated outcomes, sample size (total and per BMI category), and conclusions. Any discrepancies or conflicts during the extraction process were resolved by a third reviewer (M.A.L.).
Risk of bias assessment
The Newcastle-Ottawa scale (NOS) was used to assess the methodological quality of the included studies.
Statistical analysis
Statistical analyses were performed using the netmeta package in R ver. 3.5.2 (R Core Team). All outcomes of interest were binary and relative treatment effects were expressed as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). A frequentist NMA framework was applied. Between-study variance (τ²) was estimated using the restricted maximum likelihood method. Both fixed-effect and random-effects models were generated, with interpretation focusing primarily on the random-effects results when heterogeneity (I2) exceeded 50%.
Inconsistencies within the network were assessed by comparing direct and indirect evidence using node-splitting methods. A network heat plot was used to visualize potential inconsistency patterns.7 Forest plots were generated to display the network effect estimates for each comparison relative to the reference group. Pairwise comparisons among all interventions were further summarized in a league table, where each cell presents the estimated ORs and 95% CIs.
Heterogeneity was assessed using the I2 statistic and interpreted as follows: 0–40%, unimportant heterogeneity; 30% to 60%, moderate heterogeneity; 50% to 90%, substantial heterogeneity; and 75% to 100%, considerable heterogeneity. Random-effects models were used when I2 was greater than 50%. Sensitivity analyses were performed by excluding congress abstracts and studies with a high risk of bias.
Potential small-study effects were explored using comparison-adjusted funnel plots and Egger’s regression test. Multi-arm trials were incorporated into network analyses with appropriate adjustments to avoid unit-of-analysis errors. Statistical significance was set at p<0.05 for all analyses.
RESULTS
Characteristics of included studies
Of the total 4,170 citations found in the systematic search, seven met the eligibility criteria and were included in the study (Fig. 1, Supplementary Table 2). Among the included studies, there were five published articles and two conference abstracts; one study was from Japan and the others were from the United States. The total population consisted of 9,720 (56.8%) patients with obesity and 7,405 (43.2%) patients without obesity. Detailed characteristics of the included studies are presented in Table 1.8-14
Preferred Reporting Items for Systematic reviews and Meta-Analyses flow chart of study selection for inclusion in the study. ERCP, endoscopic retrograde cholangiopancreatography.
Characteristics of included studies
The evaluation of the risk of bias conducted using the NOS checklist is presented in Supplementary Table 3.8-14 Among the seven studies, two, including one published article8 and one conference abstract,9 were of poor quality, mainly due to the lack of comparability between the two groups.
PEP
The impact of body weight on the incidence of PEP was investigated in two separate analyses: (1) a meta-analysis comparing patients with obesity (n=9,720) and those without obesity (n=7,405) patients and (2) a NMA comparing five groups: underweight (n=352), normal weight (n=5,619), overweight (n=1,656), obesity (n=5,371), and morbid obesity (n=4,114). None of the studies defined protocols for PEP prophylaxis.
The meta-analysis revealed that obesity was not associated with an increased risk of PEP (OR, 1.33; 95% CI, 0.62–2.87; I2=73.5%) (Fig. 2). The comparison-adjusted funnel plot and Egger's test (p=0.252) revealed no small effects (Supplementary Fig. 1). Influence analysis showed that removing none of the studies significantly altered the overall estimated OR (Supplementary Fig. 2). Additionally, a secondary analysis was performed after excluding three studies (conference abstracts and those with poor quality in the risk of bias assessment).8-10 Similarly, this meta-analysis revealed that obesity was not associated with an increased risk of PEP (OR, 0.94; 95% CI, 0.68–1.32; I2=11.6%) (Supplementary Fig. 3). The comparison-adjusted funnel plot and Egger's test (p=0.704) revealed no small effects (Supplementary Fig. 4).
Pooled estimated risk of PEP in obese and non-obese groups. ERCP, endoscopic retrograde cholangiopancreatography; PEP, post-ERCP pancreatitis; IV, inverse variance; CI, confidence interval; df, degree of freedom.
In the NMA, five BMI groups were compared in five studies,8,10-13 including 10 mixed direct-indirect comparisons (Fig. 3, Supplementary Fig. 5). The analysis demonstrated significant heterogeneity (I2=64.9%) and network inconsistency (p=0.001) (Supplementary Fig. 6). Figure 4 demonstrates the odds of PEP for different BMI groups, with morbid obesity vs. obesity having the highest odds of PEP in direct (OR, 2.16; 95% CI, 0.85–5.53) and morbid obesity vs. overweight having the highest odds of PEP in network analysis (OR, 1.44; 95% CI, 0.69–3.03); however, none of the five BMI groups showed a significantly higher risk of PEP compared to that of the other groups, both in direct and network analysis. The comparison-adjusted funnel plot and Egger's test (p=0.0001) revealed a significant small-study effect (Supplementary Fig. 7).
Network plot of the included studies in PEP analysis. BMI, body mass index; ERCP, endoscopic retrograde cholangiopancreatography; PEP, post-ERCP pancreatitis.
League table with odds ratios estimates of each pair of interventions and 95% confidence intervals (CIs) based on the body mass index (BMI) group. The upper triangle shows the results from direct comparisons (row vs. column); the lower triangle shows the results from network meta-analysis (column vs. row). NA, not applicable; ERCP, endoscopic retrograde cholangiopancreatography; PEP, post-ERCP pancreatitis.
Owing to the significant heterogeneity and network inconsistency observed in the previous NMA, a secondary analysis was performed after excluding two studies (one conference abstract and one with poor-quality risk of bias assessment).8,10,15 In this NMA, five BMI groups were compared in three studies, including one direct and nine mixed direct-indirect comparisons (Fig. 5, Supplementary Fig. 8). The analysis demonstrated no significant heterogeneity (I2=35.7%) or network inconsistencies (p=0.168) (Supplementary Fig. 9). Figure 6 demonstrates the odds of PEP for different BMI groups, with morbid obesity vs. obesity having the highest odds of PEP in direct (OR, 1.66; 95% CI, 0.83–3.33) and normal weight vs. morbid obesity having the highest odds of PEP in network analysis (OR, 1.18; 95% CI, 0.87–1.61); however, none of the five BMI groups showed a significantly higher risk of PEP compared to that of the other groups, both in direct and network analysis. The comparison-adjusted funnel plot and Egger's test (p=0.456) showed no significant small-study effect (Supplementary Fig. 10).
Network plot of the included studies in PEP analysis after exclusion of poor-quality studies. BMI, body mass index; ERCP, endoscopic retrograde cholangiopancreatography; PEP, post-ERCP pancreatitis
League table with odds ratios estimates of each pair of interventions and 95% confidence intervals (CIs) based on the body mass index (BMI) group after exclusion of poor-quality studies. The upper triangle shows the results from direct comparisons (row vs. column); the lower triangle shows the results from network meta-analysis (column vs. row). NA, not applicable; ERCP, endoscopic retrograde cholangiopancreatography; PEP, post-ERCP pancreatitis.
Regarding the severity of PEP, three studies compared the severity of PEP between different BMI groups using the Cotton8,12 or revised Atlanta criteria.13 None of these studies reported a significant difference in the severity of PEP between the different BMI groups.
Secondary outcomes
The secondary outcomes included adverse events other than PEP, cannulation difficulty and success, and procedure time and success.
Owing to the small number of studies, rare occurrence of the outcome and heterogeneity in definitions and criteria, a meta-analysis was not feasible for adverse events other than for PEP. Studies have investigated bleeding,8,9,11,14 perforation,8,9,11 cholangitis,14 and cholecystitis11; however, none of them reported a significant difference between obesity and non-obesity or different BMI groups.
Difficult cannulation was investigated in two studies. One study, which defined easy cannulation as up to eight cannulation attempts and moderate/difficult cannulation as more than eight cannulation attempts, reported moderate/difficult cannulation in 10 (9.3%), 37 (11.1%), 37 (13%), and 30 (12.5%) underweight, normal weight, overweight, and patients with obesity, respectively (p=0.73).12 Another study, which did not clearly define difficult cannulation criteria, reported difficult cannulation in eight (10%), 32 (9%), seven (16%), and one (5%) underweight, normal weight, overweight, and patients with obesity, respectively (p=0.567).8 Additionally, two studies compared the rate of successful biliary cannulation between obesity and non- obesity groups: one reported successful biliary cannulation in all patients (p=1.00),14 whereas the other reported a significantly lower rate of successful biliary cannulation in patients with obesity than in those without obesity (81.5% vs. 94%, p=0.01).9
The total procedure time was investigated in two studies. Fujisawa et al.8 reported a longer cannulation time in underweight patients (50.8, 45.5, 47.8, and 39.5 minutes for underweight, normal weight, overweight, and patients with obesity, respectively); however, this was not statistically significant (p=0.123). However, a study by Zivari et al.10 reported a significantly longer procedure time in underweight patients than in patients with obesity (45.1 vs. 35.2 minutes, p=0.046).
Procedural success was evaluated in two studies. Fujisawa et al.8 reported successful treatment in 72 (94%), 351 (97%), 121 (98%), and 19 (95%) underweight, normal weight, overweight, and patients with obesity, respectively (p=0.411). Jaruvongvanich et al.14 set the rate of stone clearance in patients with choledocholithiasis as the primary outcome and reported no statistically significant difference in the rate of incomplete stone extraction between the obesity (11.3%) and non-obesity (9%) groups (p=0.51).
DISCUSSION
In this systematic review and NMA, we investigated the impact of obesity and BMI on ERCP outcomes. The results of the NMA revealed that none of the BMI groups was associated with an increased risk of PEP. Moreover, no significant differences were observed in the rates of difficult cannulation, cannulation success, stone clearance, and procedural success between the obesity and non-obesity groups.
The relationship between obesity and acute pancreatitis has been extensively studied. Reportedly, obesity increases not only the risk of development but also the complications and mortality of acute pancreatitis.16,17 Recent retrospective analyses from China revealed that BMI is a significant predictor of the onset of both acute necrotizing and edematous pancreatitis. Furthermore, the study identified that increased BMI was correlated with greater short-term recurrence rates, establishing obesity as a risk factor for the progression of pancreatitis. Notably, these findings also point to the potential adverse outcomes associated with an underweight status.18 The suggested mechanism involves adiponectin, an anti-inflammatory adipokine that protects against pancreatitis by inhibiting pro-inflammatory signaling and nuclear factor-κB.19,20 Reduced obesity levels are inversely correlated with pancreatitis severity.21 However, the association between BMI and pancreatitis may differ in the context of ERCP. This discrepancy may be due to the differences in the pathophysiology and etiology of acute pancreatitis and PEP. It has been proposed that mechanical trauma to the papilla from instrumentation, sphincterotomy, repetitive and aggressive cannulation attempts, or contrast injection can induce papillary spasm or edema. This condition is believed to elevate ductal pressure and hinder pancreatic duct drainage, triggering an inflammatory response that facilitates the intraluminal activation of proteolytic enzymes, ultimately leading to the autodigestion of pancreatic tissue and the development of PEP.22
This study found no significant association between BMI and PEP. Initially, Cotton et al.,23 in a multivariate analysis of 11,497 patients who underwent ERCP between 1994–2006, reported that obesity is significantly associated with severe PEP (OR, 5.18; 95% CI, 1.74–15.43). However, obesity was not the focus of this study, and the authors evaluated its impact on PEP using regression analysis. Among the studies included in this meta-analysis, only one study reported a significantly higher incidence of PEP in patients with obesity than in other BMI groups.8 However, the obesity group comprised only a small proportion of the study population (3.4%). Moreover, there were significant differences in age and sex between the BMI groups. Female sex and younger age, which are significant risk factors for PEP, were more common in the obesity group. Owing to the non-comparability of the groups, this study also received a poor quality rating in the risk of bias assessment, and after excluding it from the network, the inconsistency and heterogeneity of the analysis significantly improved. In general, it can be concluded that obesity and a higher BMI are not independent risk factors for PEP, and the associations reported in several studies can be attributed to methodological issues. However, a definitive conclusion is constrained by the inability to adjust for well-established PEP risk factors such as female sex, younger age, difficult cannulation, and the PEP prophylaxis protocol, as the included studies lacked the necessary stratified data. This highlights the need for future large-scale studies aimed specifically at evaluating the independent role of BMI, while controlling for fundamental confounders.
In addition to PEP, cannulation difficulty and success, stone clearance, and procedural success were investigated as secondary outcomes. However, owing to the small number of studies, low occurrence of the outcomes, and discrepancies in definitions and criteria, a meta-analysis was not feasible for these secondary outcomes. Nonetheless, a systematic review suggested that there were no significant differences in the rates of difficult cannulation, cannulation success, complete stone extraction, and procedural success between BMI groups. However, these findings should be interpreted with caution, as the lack of statistical differences could be related to the small sample size and methodological issues of the studies. Further large-scale studies with standardized and uniform outcomes are required to shed light on this topic.
Another important issue related to ERCP in patients with obesity is radiation exposure and fluoroscopy time. The technical challenges posed by ERCP in patients with obesity often lead to longer fluoroscopy times, and therefore, increased radiation exposure for patients and endoscopists. However, none of the included studies investigated fluoroscopy time or radiation exposure as outcomes. Nevertheless, an interesting study by Oh et al.24 investigated the factors associated with increased fluoroscopy time during ERCP. Through a multivariate analysis of 780 patients who underwent ERCP between 2014 and 2016, they reported that a BMI greater than 27.5 kg/m² (+4.1 minutes, 95% CI, 2.56–5.63) was associated with longer fluoroscopy time. Mechanical lithotripsy, malignant biliary obstruction, and use of a needle knife were other factors associated with prolonged fluoroscopy time. Although increased BMI may not be associated with an increased risk of adverse events, unsuccessful biliary cannulation, or procedural failure, it may be associated with increased radiation exposure, which can lead to potential health issues.
To the best of our knowledge, this is the first meta-analysis to evaluate the effects of BMI on ERCP outcomes. Categorizing patients based on BMI for NMA was one of the strengths of this study. However, this study has some limitations. First, the small number of studies included in the NMA reduces the robustness of our results. This was more notable in the evaluation of secondary outcomes, in which the small number of studies made a meta-analysis impossible. Second, BMI categorization varied among the included studies. One of the studies included in the NMA12 categorized patients into four groups: underweight, normal weight, overweight, and obesity. It is possible that a small number of patients in the obese group actually belonged to the morbid obesity group, which was examined using NMA. Furthermore, the outcome definitions, particularly for PEP and its severity, were not uniform across studies. This variability in the definitions and BMI classifications may have slightly affected the robustness of the results. The significant heterogeneity and inconsistency observed in the NMA is a limitation, although this issue was resolved after the exclusion of studies of poor quality. Finally, all included studies had a retrospective cohort design, which could introduce biases associated with cohort studies. Future large-scale prospective studies with standardized BMI categorizations, uniform outcome definitions, and strict methodologies are thus warranted to validate these findings.
In conclusion, there was no significant association between BMI and incidence or severity of PEP. Moreover, patients with obesity had similar rates of difficult cannulation, cannulation success, complete stone extraction, and procedural success compared to those of patients without obesity. Further prospective, large-scale studies are needed to shed light on this topic.
Supplementary Material
Supplementary Table 1. Search strategy in source database.
Supplementary Table 2. Excluded studies in full-text review.
Supplementary Table 3. Quality of the included studies on the basis of the Newcastle-Ottawa scale checklist.
Supplementary Fig. 1. Comparison-adjusted funnel plot of PEP in obese and non-obese groups.
Supplementary Fig. 2. Sensitivity analysis of PEP in obese and non-obese groups.
Supplementary Fig. 3. Pooled estimated risk of PEP in obese and non-obese groups after exclusion of poor-quality studies.
Supplementary Fig. 4. Comparison-adjusted funnel plot of PEP in obese and non-obese groups after exclusion of poor-quality studies.
Supplementary Fig. 5. Direct and indirect evidence proportion of PEP in overall network.
Supplementary Fig. 6. Heat plot of PEP in overall network.
Supplementary Fig. 7. Comparison-adjusted funnel plot of PEP in overall network.
Supplementary Fig. 8. Direct and indirect evidence proportion after exclusion of poor-quality studies from the network.
Supplementary Fig. 9. Heat plot of PEP in overall network after exclusion of poor-quality studies from the network.
Supplementary Fig. 10. Comparison-adjusted funnel plot of PEP in overall network after exclusion of poor-quality studies.
Supplementary materials related to this article can be found online at https://doi.org/10.5946/ce.2025.255.
Notes
Ethical Statements
Not applicable.
Conflicts of Interest
The authors have no potential conflicts of interest.
Funding
None.
Author Contributions
Conceptualization: EA, MAL, AS, MA; Data curation: EA, SK, MAL; Investigation: EA; Methodology: EA, MAL; Visualization: EA, SK, ML; Writing–original draft: EA, SK, MAL; Writing–review & editing: all authors.