Background

Acute pancreatitis is the most common major and severe complication of diagnostic and therapeutic ERCP, with the reported incidence ranging from 1.8% to 7.2% in large prospective series of nonselected patients [1–4]. The severity of post-ERCP pancreatitis can range from a mild course with one or two days prolonged hospitalization and full recovery to a devastating illness with hemorrhagic pancreatitis, pancreatic necrosis, multiorgan failure, and even death [5]. Because post-ERCP pancreatitis is predictable and possibly preventable, numerous attempts have been made to reduce the incidence and limit the severity of this complication. However, most of them have largely been disappointing.

Although the pathogenesis of post-ERCP pancreatitis is not clearly understood, a number of studies have demonstrated that an early step in the pathogenesis of acute pancreatitis is capillary endothelial injury manifested by an increase in capillary permeability [6, 7]. Subsequent researches have suggested that this capillary injury may be mediated by oxygen-derived free radicals [8, 9]. The manifestations of pancreatitis in experimental animal model can be ameliorated by blocking the action of oxygen-derived free radicals [10–12]. Based on the aforementioned findings, the idea of antioxidant supplementation for the prevention of post-ERCP pancreatitis seems rational and reasonable.

Two meta-analyses regarding allopurinol (an inhibitor of oxygen-derived free-radicals) on the prevention of post-ERCP pancreatitis have been published [13, 14]. Both of them showed that allopurinol has no significant effect on the prevention of post-ERCP pancreatitis. However, one was only based on 4 trials and included relatively modest sample sizes [13], and another included half of the studies published in the abstract form, without access to the full data [14]. Moreover, the role of other antioxidants (such as N-acetylcysteine, β-carotene) has not been well established. Recently, several relevant randomized controlled trials (RCTs) regarding prophylactic antioxidant supplementation in preventing post-ERCP pancreatitis have been published. These reports were well-performed RCTs and included an additional more than 1,200 patients. We therefore undertook a meta-analysis of randomized controlled trials to evaluate the effect of prophylactic antioxidant supplementation compared with control on the incidence and the severity of post-ERCP pancreatitis.

Methods

Literature search and inclusion criteria

Literature searches of the PubMed and Embase databases (up to May 2012) were performed to identify RCTs that compared antioxidant versus control for the prevention of post-ERCP pancreatitis. The initial search terms were antioxidant and pancreatitis, filtered by Humans andRandomized Controlled Trial. In addition, the reference lists of identified studies were manually checked to include other potentially eligible trials. This process was performed iteratively until no additional articles could be identified.

The following inclusive selection criteria were applied: (i) study design: RCT; (ii) study population: adult patients undergoing ERCP; (iii) intervention: antioxidant supplementation (no matter what type and regimen applied); (iv) comparison intervention: placebo or no intervention; and (v) outcome measure: the incidence and the severity of post-ERCP pancreatitis.

Data extraction and outcome measure

Two authors (WJG and CYW) independently extracted the following data from the selected studies: first author, year of publication, number of patients (antioxidant/control), patient characteristics, type of antioxidant, regimens of antioxidant supplementation (route, dosage, timing, frequency), study design, definition and severity of post-ERCP pancreatitis, and outcome data. Extracted data were entered into a standardized Excel file and were checked by another author (RXY). Any disagreements were resolved by discussion and consensus.

The outcome of interest was the incidence and the severity of post-ERCP pancreatitis. The definition of post-ERCP pancreatitis varied across studies, no standard definition was used in reported studies. In the majority of the studies, the definition of post-ERCP pancreatitis and the grading of its severity were based on the Cotton consensus criteria [15].

Quality assessment

The methodological quality of each trial was evaluated using the Jadad scale [16]. The scale consists of three items describing randomization (0–2 points), blinding (0–2 points), and dropouts and withdrawals (0–1 points) in the report of a randomized controlled trial. A score of 1 is given for each of the points described. A further point is obtained where the method of randomization and/or blinding is given and is appropriate; whereas it is inappropriate a point is deducted. The quality scale ranges from 0 to 5 points. Higher scores indicate better reporting. The studies are said to be of low quality if the Jadad score is ≤ 2 and high quality if the score is ≥ 3 [17].

Statistical analyses

All outcomes were expressed as RR with 95% CI. The Cochrane Q x 2 test was used to detect heterogeneity of the effects, significant heterogeneity was defined as a P value of <0.05. A fixed-effects model or random-effects model was used, depending on the absence or presence of heterogeneity. I 2 statistic was estimated to describe the percentage of the variability attributable to heterogeneity rather than sampling error. Studies with an I 2 statistic of < 25% are considered to have no heterogeneity, those with an I 2 statistic of 25% to 50% are considered to have low heterogeneity, those with an I 2 statistic of 50% to 75% are considered to have moderate heterogeneity, and those with an I 2 statistic of > 75% are considered to have high heterogeneity [18]. Whenever heterogeneity was present, sensitivity analyses based on sample size, study quality, and omitting one study in each turn were carried out to identify potential sources. We also investigated the influence of a single study on the overall pooled estimate by omitting one study in each turn.

Potential publication bias was assessed by visually inspecting of the Begg funnel plot in which the RRs were plotted against their SEs. The presence of publication bias was also evaluated by using the Begg and Egger tests [19, 20]. A P value less than 0.05 was judged as statistically significant, except where otherwise specified. All statistical analyses were performed using STATA version 11.0 (Stata Corporation, College Station, Texas, USA).

Results

Study identification and selection

The PubMed and Embase search identified 42 and 25 potential studies, respectively. A total of 67 RCTs were identified by the initial database search. Thirteen RCTs were excluded because of duplicate studies and 43 RCTs were excluded based on the titles and abstracts (reviews, nonrandomized studies, or not relevant to our analysis). The remaining 11 were then retrieved for full text review. Finally, eleven RCTs met inclusion criteria and were included in the analysis [21–31].

Study characteristics

The main characteristics of eleven RCTs included in this meta-analysis are presented in Table 1 and the definition and severity of post-ERCP pancreatitis of each included trial are described in Table 2. These studies were published between 1999 and 2011. The size of the RCT ranged from 40 to 701 (total 3,010). Among the 11 studies included here, all reported post-ERCP pancreatitis events [21–31], 8 reported mild and moderate post-ERCP pancreatitis events [22–26, 28, 29, 31], and 6 reported severe post-ERCP pancreatitis events [22–24, 26, 28, 29]. The median Jadad score of the studies included was 3 (range from 2 to 5).

Table 1 Main characteristics of randomized controlled trials included in the meta-analysis
Full size table
Table 2 Definition and severity of post-ERCP pancreatitis
Full size table

All these patients were older than 18 years and scheduled for ERCP. The selected trials used different types of antioxidant, including sodium selenite [21], allopurinol [22, 25, 26, 29–31], N-acetylcysteine [24, 27], β-carotene [23], and pentoxifylline [28]. These antioxidants were administered orally or intravenously by different regimens and formulations. Two studies used an intravenous route to administer antioxidant [21, 24], and the remaining nine studies orally applied antioxidant during the perioperative period [22, 23, 25–31]. Dosage, timing, and frequency of these antioxidants are various.

The incidence of post-ERCP pancreatitis

The outcome data of each included trial are described in Table 3. A total of 3,010 patients were included in the ten trials comparing antioxidant with control for the prevention of post-ERCP pancreatitis (1,484 in the antioxidant group and 1,526 in the control group). Altogether, 266 patients developed post-ERCP pancreatitis, 127 in the antioxidant group and 139 in the control group. Antioxidant supplementation were not associated with a significant reduction in the incidence of post-ERCP pancreatitis (RR, 0.92; 95% CI, 0.65-1.32; P = 0.665), with significant heterogeneity among the studies (I 2 = 44.7%; P = 0.054). Furthermore, when trials were divided by the type of antioxidant, there was also no significant decrease in post-ERCP pancreatitis (RR for trials with allopurinol: 0.76; 95% CI: 0.41-1.42; P = 0.396; and RR for trials with other antioxidants: 1.11; 95% CI: 0.74-1.66; P = 0.622) (Figure 1).

Table 3 Outcome data of randomized controlled trials included in the meta-analysis
Full size table
Figure 1
figure 1

Forest plot showing the effect of antioxidant supplementation on the incidence of post-ERCP pancreatitis.

Full size image

Subsequently, we performed sensitivity analyses to explore potential source of heterogeneity. Exclusion of two trials which had a modest sample size (N ≤ 100) yielded similar results (RR 0.92; 95% CI 0.62-1.39; P = 0.702), with moderate heterogeneity (I 2 = 55.2%; P = 0.022) [21, 31]. Exclusion of two studies with low quality (Jadad score ≤ 2) did not change the pooled results substantially (RR 0.96; 95% CI 0.64-1.43; P = 0.823), yet heterogeneity was still present (I 2 = 53.6%; P = 0.028) [21, 27]. Exclusion of one study conducted by Katsinelos et al. changed the overall estimate little (RR, 1.08; 95% CI, 0.85–1.37; P = 0.531), but no evidence of heterogeneity was observed among the remaining studies (I 2 = 0%; P = 0.651) [25]. Further exclusion of any single study also did not materially alter the overall combined RR (data not shown).

The severity of post-ERCP pancreatitis

We also performed meta-analyses according to the grade to explore the effect of antioxidant on the severity of post-ERCP pancreatitis. Antioxidant supplementation had no impact on mild post-ERCP pancreatitis (eight RCTs, RR 1.25, 95% CI 0.90-1.72; P = 0.183; I 2 = 0%; heterogeneity P = 0.776) (Figure 2), moderate post-ERCP pancreatitis (eight RCTs, RR 0.80, 95% CI 0.53-1.22; P = 0.304; I 2 = 0.3%; heterogeneity P = 0.426) (Figure 3), or severe post-ERCP pancreatitis (six RCTs, RR 0.71, 95% CI 0.30-1.73; P = 0.455; I 2 = 0%; heterogeneity P = 0.600) (Figure 4).

Figure 2
figure 2

Forest plot showing the effect of antioxidant supplementation on the incidence of mild post-ERCP pancreatitis.

Full size image
Figure 3
figure 3

Forest plot showing the effect of antioxidant supplementation on the incidence of moderate post-ERCP pancreatitis.

Full size image
Figure 4
figure 4

Forest plot showing the effect of antioxidant supplementation on the incidence of severe post-ERCP pancreatitis.

Full size image

Publication bias

Assessment of publication bias using Egger’s and Begg’s tests showed that there was no potential publication bias among the included trials (Egger’s test, P = 0. 443; Begg’s test, P = 0.533, Figure 5).

Figure 5
figure 5

Tests for publication bias for RR of the incidence the incidence of post-ERCP pancreatitis.

Full size image

Discussion

This is a further meta-analysis to evaluate the effect of prophylactic antioxidant supplementation on the prevention of post-ERCP pancreatitis. The present meta-analysis of ten randomized controlled trials showed that antioxidant supplementation did not significantly decrease the incidence of post-ERCP pancreatitis. In addition, antioxidant supplementation also showed no beneficial effect on the severity of post-ERCP pancreatitis.

There have been two published meta-analyses of allopurinol (an inhibitor of oxygen-derived free-radicals) for post-ERCP pancreatitis prevention [13, 14]. Both of them showed that the use of allopurinol was not associated with reduction in the incidence of post-ERCP pancreatitis. Our meta-analysis expands on these two earlier meta-analyses to provide a better characterization of the evidence base for antioxidant supplementation in preventing post-ERCP pancreatitis. First, in our analysis, there are more enlarged sample sizes than the previous analysis, giving greater power to evaluate this effect. Second, we were more capable of evaluating the effects of other antioxidants (such as N-acetylcysteine, β-carotene, and selenite) on post-ERCP pancreatitis prevention. Furthermore, we also were able to evaluate the effect of antioxidant supplementation on the severity of post-ERCP pancreatitis.

Based on the previous meta-analysis, we furthermore included other seven recent RCTs [21, 23, 25, 27, 28, 30, 31]. With the added statistical power of having 2,970 cases, the present meta-analysis suggested that antioxidant supplementation did not significantly decrease the incidence of post-ERCP pancreatitis, which was in line with the previous meta-analysis. Moreover, exclusion of any single study and sensitivity analyses based on various exclusion criteria did not materially alter the pooled results, which adds robustness to our main finding. We also assessed the effect of antioxidant supplementation on the severity of post-ERCP pancreatitis, but failed to find significant alteration.

There was significant heterogeneity between studies in the overall analysis, which was not surprising given the differences in characteristics of populations, antioxidant supplementation, and definitions of pancreatitis. Our sensitivity analyses suggest that one trial conducted by Katsinelos et al. probably contributed to the heterogeneity [25]. For this study, only 43 patients in the allopurinol group underwent biliary sphincterotomy vs. 87 in the placebo group; thus, there was significant difference in the proportions of patients with biliary sphincterotomy between the two groups (P < 0.001). Besides, exclusion of the trial in our meta-analysis would not change our result; antioxidant supplementation still did not significantly decrease the incidence of post-ERCP pancreatitis (2,767 patients; fixed-effects model: RR, 1.08; 95% CI, 0.85–1.37; P = 0.531; data from nine trials) [21–23, 25–31].

Furthermore, antioxidant therapy failed to prevent the onset of post-ERCP pancreatitis in almost all trials. Only one clinical trial in which 600 mg of allopurinol was administered twice before ERCP showed a significant decrease in the rate of post-ERCP pancreatitis. The positive result may be attributed to the high doses of allopurinol. Wisner and Renner found that only high doses of allopurinol were effective in preventing pancreatic edema and an increase in serum amylase in caerulein-induced pancreatitis in rats [11]. Low doses of allopurinol were not effective in fully suppressing xanthine oxidase, causing a “leakage” of excessive radicals.

Most of the included RCTs did not reported adverse effects of antioxidant during the study period. We found that only 1 report indicated different adverse effects of N-acetylcysteine. In the report [25], side effects most commonly attributable to N-acetylcysteine (i.e., skin rash, nausea, vomiting, and diarrhea) were observed with increased frequency in the N-acetylcysteine group (25% vs. 3.2%, N-acetylcysteine and placebo groups, respectively; P < 0.001).

The results of this meta-analysis must be interpreted cautiously in light of the strengths and limitations of the included trials. A major strength of this study is that all the included original studies used a randomized controlled design and most of them were well-performed, high-quality (Jadad score ≥ 3 in nine trials). In addition, with the enlarged sample size, we have enhanced statistical power to provide more precise and reliable effect estimates. Although the present analysis represents a complete summary of the current available evidence for antioxidant supplementation in preventing post-ERCP pancreatitis, it also serves to highlight limitations that remain. One potential limitation is the various types of antioxidant used and the lack of standard regimens of antioxidant applied in the randomized trials to date. These factors may result in the heterogeneity and have potential impact on our results. Furthermore, the variety of criteria was used to define pancreatitis. In the majority of the studies, the definition of post-ERCP pancreatitis and the grading of its severity were based on the Cotton consensus criteria [15]. Finally, exploration of the impact of antioxidant supplementation on other clinically meaningful endpoints including hyperamylasemia and length of the hospital stay has not been sufficient because of sparse and inconsistent reporting across the reviewed studies.

Conclusions

In conclusion, the results suggest that antioxidant supplementation shows no beneficial effect on the incidence and the severity of post-ERCP pancreatitis. There is currently a lack of evidence to support using antioxidants for the prevention of post-ERCP pancreatitis.