Efficacy of indomethacin for the prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis: A comprehensive meta-analysis of randomized controlled trials

Abstract

BACKGROUND

Post-endoscopic retrograde cholangiopancreatography pancreatitis (PEP) is a prevalent and potentially serious complication in patients undergoing endoscopic retrograde cholangiopancreatography.

AIM

To comprehensively assess the efficacy of indomethacin therapy in reducing PEP risk.

METHODS

We searched PubMed, EMBASE, Scopus, and Cochrane Library databases to identify randomized controlled trials (RCTs) that compared rectal indomethacin with a control group to prevent PEP. Duplicates were removed, and studies were included based on the established inclusion criteria. We used the Cochrane Collaboration’s tool to assess the risk of bias in the RCTs. A random-effects model was applied to produce pooled risk ratios (RRs) with 95% confidence intervals (CIs).

RESULTS

We included a total of 30 RCTs involving 16977 patients. Compared to the control group, rectal indomethacin showed comparable rates of overall PEP (PEP; RR = 0.85, 95%CI: 0.69-1.04, I² = 79%) with no statistically significant difference of RR in mild (RR = 0.92, 95%CI: 0.74-1.14), moderate (RR = 0.78, 95%CI: 0.59-1.02), or severe PEP (RR = 1.12, 95%CI: 0.75-1.67). There was also no difference in cases of adverse events (RR = 0.97, 95%CI: 0.69-1.35), abdominal pain (RR = 1.14, 95%CI: 0.80-1.62), bleeding (RR = 1.07, 95%CI: 0.70-1.63), or mortality (RR = 0.86, 95%CI: 0.56-1.33) between the two groups. Subgroup analyses were also performed.

CONCLUSION

Rectal indomethacin appears to be safe and may offer benefit in selected high-risk patients, though findings should be interpreted with caution due to high heterogeneity.

Core Tip: This meta-analysis assessed the efficacy of rectal indomethacin in the prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis by reviewing the results of 30 randomized controlled trials. Indomethacin didn’t show a significant reduction in overall post-endoscopic retrograde cholangiopancreatography pancreatitis rates or adverse events relative to controls, however, it may be beneficial in particular high-risk patients, but with caution given to significant heterogeneity in the results.

INTRODUCTION

Endoscopic retrograde cholangiopancreatography (ERCP) is a frequently used medical procedure for diagnosing and treating pancreaticobiliary diseases[1,2]. However, post-ERCP pancreatitis (PEP) remains a prevalent and potentially serious adverse event, occurring in 3%-10% of patients without any particular criteria and resulting in significant morbidity, prolonged hospitalization, increased health care costs, and even mortality in severe cases[3-6]. Risk factors include patient age, gender, and procedural difficulties, emphasizing the need for careful management during ERCP[7].

For several decades, the insertion of a temporary, prophylactic stent in the pancreatic duct during ERCP was the sole effective preventative measure for patients at high risk for PEP[8-10]. Given the significant consequences of PEP, there has been a tremendous amount of interest in studying pharmacologic and procedural interventions to lower its prevalence[11-15]. Rectal indomethacin is a highly promising agent because of its affordability, ease of administration, and favorable safety profile[16,17]. Research indicates that certain agents involved in inflammation, such as phospholipase A2 and proinflammatory cytokines, are important factors in the onset of acute pancreatitis, which clearly illustrates the concept behind their use[18,19].

Over the past decade, several randomized controlled trials (RCTs) have evaluated the efficacy of prophylactic indomethacin in preventing PEP, but the results have been mixed and inconclusive[20-23]. Previous meta-analyses have also reached conflicting conclusions, likely due to differences in trial inclusion criteria, quality assessment, and analytical methods[8,22,24-26]. To better understand the actual efficacy of indomethacin in PEP prophylaxis, a comprehensive review of RCTs is required, given the conflicting data and potential impact on clinical practice.

This study aims to offer a more accurate estimate of the treatment impact and investigate potential sources of variation by combining data from multiple high-quality trials. Therefore, we conducted an updated and comprehensive systematic review and meta-analysis to critically evaluate the evidence and provide the most reliable estimates of the pros and cons of indomethacin for preventing PEP.

MATERIALS AND METHODS

This meta-analysis was conducted by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement[27].

Protocol registration

We registered this study at Prospero with the number CRD42024563974. (https://www.crd.york.ac.uk/PROSPERO/view/CRD42024563974).

Search strategy

Author (Tian F) searched the PubMed/MEDLINE, EMBASE, Scopus, and Cochrane Library databases from inception to June 30, 2024 to identify relevant studies. The search strategy used keywords and medical subject headings terms related to ERCP, indomethacin, and pancreatitis, with no restrictions on language. We also manually searched the reference lists of the included studies and previous meta-analyses for additional eligible trials (search strategy in Supplementary material).

Eligibility criteria

Two reviewers (Tian F and Huang ZC) independently screened titles and abstracts and then full-text articles for eligibility. Disagreements were resolved by consensus with a third reviewer (Khizar H). Studies were included and excluded as per given criteria.

Inclusion criteria: (1) Had a RCT design; (2) Compared rectal indomethacin with a control for PEP prophylaxis; and (3) Reported the incidence of PEP.

Exclusion criteria: (1) Observational studies, case reports, and trials that used other non-steroidal anti-inflammatory drugs (NSAIDs); (2) Studies that used non-rectal routes of indomethacin administration; and (3) Duplicate publications or studies in which the required outcomes were missing.

Data extraction and quality assessment

Two reviewers (Tian F and Huang ZC) independently extracted data, including study characteristics (author, year, country, study settings, sample size, and ERCP indications), patient demographics, indomethacin dose, and outcome data, via a standardized form. Discrepancies were resolved by a rereview of the primary articles. We employed the Cochrane Collaboration’s tool for assessing risk of bias in randomized trials to evaluate the risk of bias[28]. This tool explores various domains, including random sequence generation, allocation concealment, blinding of participants and staff members, blinding of outcome evaluation, inadequate outcome data, selective reporting, and other sources of bias. We evaluated each domain and assigned a rating of low, some concern, or high risk of bias for the studies.

Outcomes

The primary outcome evaluated was the incidence of PEP, which is typically defined by the onset or worsening of abdominal discomfort that is consistent with pancreatitis, followed by an elevation in serum amylase or lipase ≥ 3 times the upper limit of normal at 24 hours post-procedure. The secondary outcomes included the severity of pancreatitis (based on criteria), length of hospital stay, and adverse events. We conducted subgroup analyses via the following comparisons: Indomethacin alone vs indomethacin combined, indomethacin combined vs control, indomethacin vs placebo, indomethacin combined vs placebo, indomethacin vs saline/saline combined, indomethacin vs glycerin/epinephrine or combined, indomethacin vs indomethacin + stent, two-arm vs multiple-arm studies, selected patients vs unselected patients, single-center vs multicenter studies, and patients aged < 60 vs > 60 years.

Statistical analysis

We pooled dichotomous outcomes using risk ratios (RRs) with 95% confidence intervals (CIs). The continuous outcomes were pooled by calculating the mean differences (MDs) along with their 95%CIs. A random effects model with the Mantel-Haenszel method was applied to evaluate the variation among studies. We pooled the multivariance analysis outcomes of the studies via the generic inverse variance and random effect methods to compute the pooled RRs and 95%CIs. We evaluated heterogeneity via the Cochran Q test and calculated the I² statistic. The heterogeneity was classified as low, moderate, or high on the basis of I² values of 30%-49%, 50%-74%, and more than 75%, respectively[29]. Whereas less than 30 was regarded as negligible.

Assessing publication bias involved visually inspecting funnel plots and conducting Egger’s regression test[30]. We evaluated to determine the impact of individual studies on the pooled estimate. This was done via sensitivity analysis via the leave-one-out method. We conducted subgroup analyses to examine various patient and procedural factors. P ≤ 0.05 was considered statistically significant. We conducted our analyses via Review Manager 5.4 (Cochrane Collaboration).

RESULTS

Selection process

The literature search yielded 479 records, of which the meta-analysis included 30 RCTs with a total of 16977 patients[20,31-38] (Figure 1).

Figure 1.

PRISMA flow diagram of study selection.

Study characteristics

The characteristics of the studies are presented in Tables 1 and 2. The mean age varied between 42 and 66 years, with female participants constituting 29% to 100% of the total. The primary indications for ERCP were choledocholithiasis and other standard reasons for the procedure. Most trials used a single 100 mg rectal dose; exceptions included 200 mg escalation (Fogel et al[36], 2020) and split dosing (Lai et al[37], 2019), either alone or in combination, before or immediately after the procedure. These studies were conducted worldwide, with seven of them conducted in China, six in the United States, seven in Iran, two in Mexico, two in India, two in Hungary and other different countries. Seven studies were conducted as multiple-arm RCTs[32,39-44]. We analyzed the data from all these studies for the indomethacin group, which were classified as arms (a, b, c, d). Consequently, we analyzed a total of 50 double-arm comparisons to determine the primary outcome. Ten studies used indomethacin in conjunction with other medications, whereas twenty studies used it as a single therapy in comparison with a control group. Seventeen studies included all eligible patients receiving ERCP for many different indications, whereas thirteen trials included selected patients with moderate to high risk of PEP.

Table 1.

Characteristics of included studies and outcomes

Ref.	Setting	Intervention	Control	Indication	Patients	Age	Female	PEP
								Overall	Mild	Moderate	Severe
Romano-Munive et al[50], 2021, Mexico	Multicenter, double blinded, unselected patients	100 mg of rectal indomethacin with 10 mL water	100 mg of rectal indomethacin with 1:10000 epinephrine dilution (0.1 mg/mL)	Naïve papilla and indication for ERCP	EI = 275	50.3 ± 21.4	188	10	8	1	1
					WI = 273	51.8 ± 20.3	192	14	9	5	0
Patai et al[22], 2017, Hungary	Single-center, double blinded, unselected patients	100 mg of rectal indomethacin	Placebo	Intact papilla undergoing biliary endoscopic therapy	Indomethacin = 270	66.25	181	18	15	2	1
					Placebo = 269	64.51	181	37	33	3	1
Elmunzer et al[35], 2012, United States	Multicenter, single blinded, selected patients	100 mg of rectal indomethacin suppositories	Placebo	Risk of post-ERCP pancreatitis	Indomethacin = 295	44.4 ± 13.5	229	27	14	13	3
					Placebo = 307	46.0 ± 13.1	247	52	25	27	3
Levenick et al[51], 2016, United States	Single-center, single blinded, unselected patients	100 mg of rectal indomethacin suppositories	Placebo	Patients undergoing ERCP	Indomethacin = 223	64.9	118	16	16	0	0
					Placebo = 226	64.3	118	11	9	1	1
Liu et al[52], 2024, China	Single-center, single blinded, unselected patients	100 mg of rectal indomethacin suppositories	Placebo	Patients undergoing ERCP for bile duct stones	Indomethacin = 58	61.6 ± 15.6	29	4	3	1	0
					Placebo = 109	62.9 ± 15.2	49	22	11	9	2
Li et al[53], 2019, China	Single-center, single blinded, unselected patients	100 mg indomethacin suppositories	Glycerin suppository	Patients undergoing ERCP for bile duct stones	Indomethacin = 50	55.68 ± 13.58	31	6	NA	NA	NA
					Glycerin = 50	58.70 ± 13.60	37	16
Mohammad Alizadeh et al[44], 2017, Iran	Single-center, single blinded, unselected patients	100 mg indomethacin	100 mg diclofenac and 500 mg naproxen	Patients undergoing ERCP	Indomethacin = 122	58.0 ± 16.8	65	7	3	2	2
					Diclofenac = 124	56.5 ± 18.7	66	5	2	1	2
					Naproxen = 126	54.8 ± 13.7	66	20	7	8	5
Alavinejad et al[32], 2022, multiple countries	Multicenter, single blinded, unselected patients	100 mg indomethacin	1200 mg oral N-acetyl cysteine, N-acetyl cysteine plus indomethacin, placebo	Standard indications for ERCP	Indomethacin = 138	55.36	85	24	15	6	3
					N-acetyl cysteine = 84	57.44	49	9	2	7	0
					N-acetyl cysteine + indomethacin = 115	56.11	62	9	3	6	0
					Placebo = 95	61.53	55	19	10	7	2
Guha et al[38], 2023, India	Single-center, single blinded, unselected patients	100 mg indomethacin	Vigorous hydration	Standard indications for ERCP	Indomethacin = 174	43.7 ± 14.4	119	5	NA	5	3
					Aggressive hydration = 178	44.3 ± 14.7	127	1		1	0
Sotoudehmanesh et al[23], 2007, Iran	Single-center, double blinded, unselected patients	100 mg indomethacin	Placebo	Standard indications for ERCP	Indomethacin = 245	58.4 ± 17.1	134	7	NA	NA	NA
					Placebo = 245	58.1 ± 16.8	130	15
Andrade-Dávila et al[33], 2015, Mexico	Single-center, single blinded, selected patients	100 mg indomethacin	Glycerin suppository	Risk of post-ERCP pancreatitis	Indomethacin = 82	51.59 ± 18.55	51	4	3	1
					Glycerin = 84	54.0 ± 17.85	59	17	14	3
Qian et al[54], 2022, China	Single-center, double blinded, unselected patients	100 mg indomethacin	Glycerin suppository	Risk of post-ESWL pancreatitis	Indomethacin = 685	46 (35-54)	194	60	59	1	0
					Glycerin = 685	47 (37-54)	197	84	79	5	0
Wu et al[43], 2023, China	Single-center, single blinded, selected patients	100 mg indomethacin	0.25 mg of somatostatin indomethacin + somatostatin placebo	Risk of post-ERCP pancreatitis	Indomethacin = 366	52.7 ± 14.2	196	76	NA	NA	NA
					Somatostatin = 424	49.3 ± 16.6	163	22
					Somatostatin + Indomethacin = 420	51.9 ± 16.2	196	22
					Placebo = 248	50.4 ± 15.9	116	48
Döbrönte et al[34], 2014, Hungary	Multicenter, single blinded, unselected patients	100 mg indomethacin	Placebo	Risk of post-ERCP pancreatitis	Indomethacin = 347	NA	214	20	16	4	NA
					Placebo = 318		212	22	18	4
Fogel et al[36], 2020, United States	Multicenter, double blinded, selected patients	100 mg indomethacin	200 mg indomethacin	High risk of post-ERCP pancreatitis	Indomethacin = 515	49.3 (15.2)	392	76	48	28	NA
					Placebo = 522	50.4 (15)	421	65	37	28
Norouzi et al[55], 2023, Iran	Single-center, double blinded, selected patients	100 mg indomethacin + somatostatin	100 mg indomethacin + saline	High risk of post-ERCP pancreatitis	Indomethacin = 192	63.03 (16.57)	116	22	9	11
					Control group = 184	62.95 (15.58)	98	28	12	12
Sadeghi et al[56], 2023, Iran	Single-center, unselected patients	100 mg indomethacin	100mg indomethacin + vitamin C	Standard indications for ERCP	Indomethacin = 165	59.0 ± 14.4	96	27	19	4	4
					Control group = 165	62.0 ± 14.1	93	17	14	2	1
Kamal et al[57], 2019, United States, India	Multicenter, double blinded, selected patients	100 mg indomethacin	100 mg indomethacin + 20 mL of 0.02% epinephrine	High risk of post-ERCP pancreatitis	Indomethacin = 482	52.16 (14.3)	275	31	NA	NA	4
					Control group = 477	52.56 (15.6)	276	32			7
Elmunzer et al[20], 2024, Canada	Multicenter, double blinded, selected patients	100 mg indomethacin	100 mg indomethacin + stent	High risk of post-ERCP pancreatitis	Indomethacin = 975	55.6 (16.4)	599	145	67	58	20
					Control group = 975	55.8 (16.3)	596	110	52	44	14
Makhzangy et al[21], 2022, Egypt	Single-center, unselected patients	100 mg indomethacin	100 mg indomethacin + saline	Standard indications for ERCP	Indomethacin = 60	45.27 ± 15.39	29	5	NA	NA	NA
					Control group = 60	42.3 ± 14.28	37	0
Luo et al[58], 2019, China	Multicenter, double blinded, unselected patients	100 mg indomethacin + epinephrine	100 mg indomethacin + saline	Standard indications for ERCP	Indomethacin = 576	62 (50-71)	281	49	45	4	NA
					Control group = 582	61 (49-71)	280	31	29	2
Hosseini et al[40], 2016, Iran	Single-center, double blinded, selected patients	100 mg indomethacin	3 L normal saline. Indomethacin + normal saline 2 g glycerin suppositories	Standard indications for ERCP for CBD	Indomethacin = 100	51.20 ± 12.12	40	11	NA	NA	NA
					Intravenous saline = 100	50.76 ± 13.32	53	10
					Normal saline + indomethacin = 101	47.91 ± 11.06	62	0
					Glycerin = 105	49 ± 14.26	49	17
Abdi et al[31], 2024, Iran	Single-center, double blinded, unselected patients	100 mg indomethacin plus CoQ10	Indomethacin plus placebo	Standard indications for ERCP	Indomethacin = 166	55.0 ± 13.1	92	17	14	2	1
					Placebo = 166	58.0 ± 13.4	87	25	19	3	3
Mok et al[41], 2017, United States	Single-center, double blinded, selected patients	100 mg Indomethacin plus normal saline	Normal saline + placebo LR + placebo LR + India	High risk of post-ERCP pancreatitis	Indomethacin + normal saline = 48	62	33	6	NA	NA	1
					Normal saline + placebo = 48	58	29	10			0
					LR + placebo = 48	58	35	9			0
					LR + indomethacin = 48	63	23	3			0
Sotoudehmanesh et al[59], 2014, Iran	Single-center, double blinded, selected patients	100 mg indomethacin + dinitrate tablet	100 mg indomethacin + placebo	Standard indications for ERCP	Indomethacin = 150	58.4 ± 17.8	74	10	NA	NA	NA
					Placebo = 150	58.6 ± 17.5	80	23
Wang et al[42], 2020, China	Single-center, double blinded, selected patients	Indomethacin + nitroglycerin	Placebo suppository PSP with placebo suppository + tablet	Standard indications for ERCP only female	Indomethacin = 176	66.87 ± 13.04	176	9	5	4	0
					Placebo = 176	63.5 ± 14.4	176	34	14	20	0
					PSP = 174	66.30 ± 12	174	21	13	8	0
Lai et al[37], 2019, Taiwan	Single-center, double blinded, unselected patients	100 mg rectal indomethacin pre-ERCP + 100 post	100 mg rectal indomethacin post-ERCP	Standard indications for ERCP	Indomethacin = 87	60.5 ± 16.9	33	4	NA	NA	NA
					Placebo = 75	59.3 ± 15.7	28	5
Sotoudehmanesh et al[60], 2019, Iran	Single-center, double blinded, unselected patients	100 mg rectal indomethacin + 5 mg isosorbide	100 mg rectal indomethacin + 5 mg isosorbide + pancreatic duct stent	Standard indications for ERCP	Indomethacin = 207	56.8 (17.2)	120	33	27	6	NA
					PSP = 207	53.9 (16.8)	131	26	22	4
Sperna Weiland et al[61], 2021, Netherlands	Multicenter, double blinded, selected patients	100 mg rectal diclofenac or indomethacin	100 mg rectal diclofenac or indomethacin + hydration	Moderate to high risk of post-ERCP pancreatitis	Indomethacin = 425	60 (49-71)	250	39	29	10	NA
					Indomethacin + hydration = 388	57 (44-71)	232	30	27	3
Hatami et al[39], 2018, Iran	Single-center, double blinded, selected patients	100 mg indomethacin	10 mL epinephrine (diluted to 1/10000 in saline) epinephrine + indomethacin	High-risk post-ERCP pancreatitis	Indomethacin = 6	58.06 ± 17.1	27	6	4	1	1
					Epinephrine = 68	59.59 ± 15.68	33	1	1	0	0
					Epinephrine + indomethacin = 58	59.62 ± 15.36	34	0	0	0	0

EI: Epinephrine and rectal indomethacin; WI: Sterile water and rectal indomethacin; LR: Lactated Ringer’s solution; PSP: Prophylactic pancreatic stent placement; NA: Not available; PEP: Post-endoscopic retrograde cholangiopancreatography-pancreatitis; ERCP: Endoscopic retrograde cholangiopancreatography; ESWL: Extracorporeal Shock Wave Lithotripsy; CBD: Common bile duct stones.

Table 2.

Outcome of the studies

Ref.	Total adverse events	Abdominal pain	Bleeding	Cholangitis	Mortality
Romano-Munive et al[50], 2021, Mexico	EI = 40	9	9	5	2
	WI = 41	5	3	10	6
Patai et al[22], 2017, Hungary	Indomethacin = 30	NA	9	2	NA
	Placebo = 42		3	2
Elmunzer et al[35], 2012, United States	Indomethacin = 4	NA	4	NA	NA
	Placebo = 9		7
Levenick et al[51], 2016, United States	Indomethacin = 4	NA	4	NA	0
	Placebo = 6		6		3
Liu et al[52], 2024, China	Indomethacin = NA	NA	NA	NA	0
	Placebo = NA				1
Guha et al[38], 2023, India	Indomethacin = 11	43	3	NA	3
	Placebo = 8	40	4		3
Andrade-Dávila et al[33], 2015, Mexico	Indomethacin = 2	NA	2	NA	NA
	Placebo = 3		3
Qian et al[54], 2022, China	Indomethacin = 65	NA	NA	5	NA
	Placebo = 99			13
Wu et al[43], 2023, China	Indomethacin = 62	NA	45	NA	2
	Somatostatin = 21		18		3
	Somatostatin + indomethacin = 25		21
	Placebo = 43		30
Fogel et al[36], 2020, United States	Indomethacin = 6	NA	6	NA	NA
	Placebo = 8		8
Kamal et al[57], 2019, United States, India	Indomethacin = 6	NA	0	NA	3
	Placebo = 8		10		3
Luo et al[58], 2019, China	Indomethacin = 42	NA		4	8
	Placebo = 44			6	9
Mok et al[41], 2017, United States	NA	NA	NA	NA	2
					1
					2
					1
Sotoudehmanesh et al[59], 2014, Iran	NA	1	NA	NA	NA
		2
Wang et al[42], 2020, China	NA	NA	4	4	NA
			6	8
			4	4
Sperna Weiland et al[61], 2021, Netherlands	Indomethacin = 30	NA	NA	NA	12
	Indomethacin + hydration = 24				11

EI: Epinephrine and rectal indomethacin; WI: Sterile water and rectal indomethacin; NA: Not available.

Quality assessment

The risk of bias assessment is presented in Supplementary Figure 29. Fifteen studies had some concern regarding risk of bias, whereas the remaining fifteen studies had a low risk of bias across all domains. Most studies raised concerns about the randomization procedures and the participant recruitment processes. The overall quality of evidence was rated as high to moderate via the Grading of Recommendations Assessment, Development, and Evaluation approach.

Primary outcome

The analysis of the included studies revealed that rectal indomethacin did not significantly reduce the risk of PEP (RR = 0.85, 95%CI: 0.69-1.04, P = 0.12, I² = 79%), despite significantly high heterogeneity (Figure 2).

Figure 2.

Forest plot of the incidence of post-endoscopic retrograde cholangiopancreatography pancreatitis. CI: Confidence intervals.

Secondary outcomes

Indomethacin did not reduce the risk of mild PEP (RR = 0.92, 95%CI: 0.74-1.14, P = 0.43, I² = 55%) or moderate PEP (RR = 0.78, 95%CI: 0.59-1.02, P = 0.07, I² = 28%), and severe PEP (RR = 1.12, 95%CI: 0.75-1.67, P = 0.57, I² = 0%) in patients undergoing ERCP (Figure 3). The adverse events RR (RR = 0.97, 95%CI: 0.69-1.35; P = 0.84; I² = 83%) was also not significantly different between the indomethacin group and the control group. Abdominal pain (RR = 1.14, 95%CI: 0.80-1.62; P = 0.47; I² = 0%) and bleeding (RR = 1.07, 95%CI: 0.70-1.63; P = 0.77; I² = 71%) also showed no differences. However, there was a significant difference in the number of patients with cholangitis (RR = 0.56, 95%CI: 0.34-0.93; P = 0.02; I² = 0%). There was also no difference in the mortality rate (RR = 0.86, 95%CI: 0.56-1.33; P = 0.51; I² = 0%) between the two groups. There was also no difference in the pooled MD of hospital stay (MD = 0.00, 95%CI: 0.03-0.02; P = 0.95; I² = 0%) between the two groups (Supplementary Figures 1-6).

Figure 3.

Forest plot of post-endoscopic retrograde cholangiopancreatography pancreatitis. A: Mild post-endoscopic retrograde cholangiopancreatography pancreatitis (PEP); B: Moderate PEP; C: Severe PEP. CI: Confidence intervals.

The results of the multivariate analysis also revealed that there was a significant difference in the number of cases of PEP between the two groups concerning patients with prior cholecystectomy (pooled RR = 0.42, 95%CI: 0.24-0.71; P = 0.001; I² = 0%), a history of pancreatitis (RR = 0.67, 95%CI: 0.47-0.95; P = 0.03, I² = 0%), difficult cannulation (RR = 0.62, 95%CI: 0.44-0.89; P = 0.009; I² = 59%), pancreatic sphincterotomy (RR = 0.29, 95%CI: 0.14-0.48; P < 0.00001; I² = 25%), biliary sphincterotomy (RR = 0.85, 95%CI: 0.74-0.99; P = 0.03; I² = 0%), and a pancreatic duct stent (RR = 1.14, 95%CI: 1.19-1.67; P < 0.0001; I² = 0%). However, there was no impact of patient age > 40 years (RR = 0.79, 95%CI: 0.59-1.05; P = 0.10; I² = 62%), female sex (RR = 0.80, 95%CI: 0.60-1.07, P = 0.13, I² = 56%), or trainee involvement (RR = 1.79, 95%CI: 0.89-3.60; P = 0.10, I² = 84%) on PEP cases (Table 3 and Supplementary Figures 7-15).

Table 3.

Results of the subgroup analysis of multivariance analysis outcomes

Outcomes	Studies	RR	95%CI	P value	Heterogeneity
Age > 40 years	13	0.79	0.59-1.05	0.10	62%
Female patients	14	0.80	0.60-1.07	0.13	56%
Prior cholecystectomy	4	0.42	0.24-0.71	< 0.01	0%
History of pancreatitis	7	0.67	0.47-0.95	0.03	0%
Difficult cannulation	11	0.62	0.44-0.89	< 0.01	59%
Pancreatic sphincterotomy	9	0.29	0.18-0.46	< 0.01	25%
Biliary sphincterotomy	7	0.85	0.74-0.99	0.03	0%
Trainee involvement	6	1.79	0.89-3.60	0.10	84%
Pancreatic duct stent	6	1.41	1.19-1.67	< 0.01	0%

RR: Risk ratios; CI: Confidence interval.

Subgroup analyses

The positive impact of indomethacin on PEP was the same in all subgroups, including those that were given indomethacin alone vs indomethacin combined, indomethacin combined vs control, indomethacin vs placebo, indomethacin combined vs placebo, indomethacin vs saline/saline combined, indomethacin vs glycerin/epinephrine or combined, indomethacin vs the indomethacin + stent, study groups with two arms vs studies with multiple arms, selected patients vs unselected patients, and patients aged < 60 vs patients aged > 60. Sensitivity analysis via the leave-one-out approach did not significantly alter the pooled estimate in subgroup analyses. However, in the overall analysis for PEP, after removing three studies (zero cases of PEP in the control group), a significant difference in the RR for PEP incidence (RR = 0.82, 95%CI: 0.66-1.00; P = 0.05) (Supplementary Tables 1 and 2, Supplementary Figures 16-21).

Publication bias

Asymmetry in the funnel plot, especially the absence of studies at the bottom center, suggests possible publication bias, in which smaller studies with non-significant results may be neglected. However, the prevalence of studies on both sides of the center line, including those with negative impacts, suggests that if a bias exists, it may not be significant. While there are indications of potential publication bias, real heterogeneity between studies may also contribute to the observed pattern; that’s why this can be neglected (Supplementary Figures 22-28). Egger’s test results also showed that there is minor risk of bias (P > 0.05).

DISCUSSION

This meta-analysis of 30 RCTs, including 16977 patients, evaluated the efficacy and safety of rectal indomethacin in preventing PEP across diverse patient groups and clinical environments. Our data indicate that rectal indomethacin may confer a protective effect against PEP. However, the aggregate findings lacked statistical significance, and there was significant heterogeneity across the studies included. These advantages were noted across patient subgroups and indications for ERCP, accompanied by an excellent safety profile.

A preliminary analysis of all the trials (50 comparisons) that were included did not reveal a significant decrease in the number of cases of PEP between the two groups (RR = 0.85, 95%CI: 0.69-1.04; P = 0.12). The heterogeneity was significantly higher (I² = 79%), reflecting considerable variety in the study populations, methodologies, and results. Following an impact analysis and the exclusion of three studies that significantly impacted the pooled estimate [due to zero PEP cases in control group in these three studies: (1) Hosseini et al[40]: Indomethacin vs indomethacin + normal saline showing 0 vs 11 cases; (2) Hatami et al[39]: Indomethacin vs indomethacin + epinephrine showing 0 vs 6 cases; and (3) Makhzangy et al[21]: Indomethacin vs control showing 0 vs 5 cases] the difference between the indomethacin and control groups reached statistical significance (47 comparisons) (RR = 0.82, 95%CI: 0.66-1.00; P = 0.05), despite the persistence of high heterogeneity (I² = 80%). This observation emphasizes the importance of assessing the potential influence of individual research on aggregate results, as well as the need for careful interpretation amidst significant variability. In the evaluation of PEP severity, rectal indomethacin did not significantly lower the probability of mild, moderate, or severe PEP. However, the overall incidence of PEP was lower than that of the control group.

We performed subgroups analyses on the basis of indomethacin alone vs indomethacin combined, indomethacin combined vs control, indomethacin vs placebo, indomethacin combined vs placebo, indomethacin vs saline/saline combined, indomethacin vs glycerin/epinephrine or combined, indomethacin vs the indomethacin + stent, study groups with two arms vs studies with multiple arms, selected patients vs unselected patients, and patients aged < 60 vs patients aged > 60. But the results also suggested that the heterogeneity was higher in most of the analysis and we were unable to find the real cause of this heterogeneity.

According to multivariate analysis, rectal indomethacin significantly reduced the risk of PEP in certain groups of patients who had a history of cholecystectomy, pancreatitis, difficult cannulation, pancreatic sphincterotomy, biliary sphincterotomy, or pancreatic duct stenting. These results indicate that rectal indomethacin may be especially advantageous for high-risk patients and patients undergoing certain endoscopic procedures. In contrast, patient age over 40 years, female sex, and trainee involvement did not significantly influence the effectiveness of rectal indomethacin in preventing PEP. Evidence of benefit is most robust in high-risk patients given 100 mg rectally immediately after ERCP. Some studies suggest pre-ERCP dosing is equally effective, while dose escalation to 200 mg provides no added benefit[3,37].

Subgroup and sensitivity analyses revealed that the beneficial impact of rectal indomethacin on PEP was consistent across diverse trial designs, comparator groups, and patient demographics. The aggregated estimates remained consistent when individual studies were successively excluded from the analysis, except for two studies that markedly affected the overall results for PEP incidence. The findings align with and enhance prior meta-analyses. Patai et al[22] reported results from 17 trials indicating that indomethacin treatment resulted in a reduction in PEP compared with that in the control group. Akshintala et al[8] reported that NSAID use may reduce pancreatitis in patients undergoing ERCP, as evidenced by 55 RCTs that applied various prophylaxis methodologies. The current study includes the largest number of trials utilizing indomethacin with patients to date, increasing the credibility of the results. We additionally present novel data on significant clinical events, including the severity of pancreatitis and duration of hospitalization, along with pooled RRs for multivariate analysis.

Rectal indomethacin should be considered alongside additional preventative measures for PEP. According to the European Society of Gastrointestinal Endoscopy/American Society for Gastrointestinal Endoscopy guidelines, patients are classified as high-risk if they exhibit at least one definite risk factor or two or more likely risk factors[18,45-47] (Supplementary Table 3). Procedural methods such as guidewire-assisted cannulation, pancreatic duct stenting, and early precut sphincterotomy have been linked to a reduced risk[18,47]. Patient selection is most important because characteristics such as sphincter Oddi dysfunction, difficult cannulation, prior cholecystectomy, pancreatic sphincterotomy, biliary sphincterotomy, and a history of pancreatitis increase vulnerability, as demonstrated by the analysis[48]. A comprehensive strategy that includes pharmacologic prevention and risk assessment may yield the most significant advantage[8].

Our data demonstrated that the combination of indomethacin exhibited superior efficacy and reduced the risk of PEP compared with the control/placebo combination. Subgroup analysis demonstrated that indomethacin outperforms glycerin/epinephrine. A meta-analysis between indomethacin and topical epinephrine for PEP prevention also stated that combination of indomethacin and topical epinephrine also showed similar results[26]. Patients with a mean age over 60 years, as well as those with a high to moderate risk of pancreatitis, showed increased treatment efficacy. Another study between indomethacin and diclofenac in the prevention of PEP also showed that both performs similarly that align with our study results[22,49].

The meta-analysis suggested potential benefits of rectal indomethacin in reducing PEP risk, especially in moderate cases. However, study heterogeneity and inconsistent results necessitate further research. The evidence is strongest in high-risk patients, but given the safety profile and major society recommendations, many centers administer rectal NSAIDs universally. Future studies should focus on large-scale trials with standardized protocols, optimal dosing, and patient subgroups. Investigating the underlying mechanisms of the protective effect of indomethacin could refine its use and guide new prevention strategies.

This meta-analysis offers multiple strengths, such as a comprehensive literature review, rigorous study selection and data extraction processes, and thorough subgroup and sensitivity analyses. However, certain limits must be acknowledged. The significant heterogeneity among the included studies may restrict the generalizability of the findings and necessitate careful interpretation. Second, the quality of the included studies was inconsistent, with some exhibiting relatively small sample sizes or being conducted in single centers, potentially introducing bias. The optimal timing, dosage, and amount of rectal indomethacin administration for PEP prevention remain unclear and necessitate additional research. Some studies applied co-interventions (e.g., aggressive hydration, prophylactic pancreatic duct stenting, nitrates, topical epinephrine) that may have influenced the pooled results.

CONCLUSION

This meta-analysis found that rectal indomethacin did not show a statistically significant difference in the incidence of PEP. But the pharmacological safety profile appears favorable, with no significant increase in adverse events compared to controls. The evidence is strongest in high-risk patients, but given the safety profile and major society recommendations, many centers administer rectal NSAIDs universally. However, careful interpretation of these results is required due to the significant heterogeneity among studies.