Endoscopic Management of Infected Necrotizing Pancreatitis: an Evidence-Based Approach

Metal stents were proposed as an alternative to the traditionally used double-pigtail stents (DPS) for endoscopic drainage of WON. DPS are limited by a small lumen (7 French) that may result in inadequate drainage. A larger diameter of metal stents would permit spontaneous passage of necrotic tissue into the stomach, improving the success rate of endoscopic drainage.

Although originally designed for biliary strictures, fully covered self-expandable metal stents (FC-SEMS) were the first metal stents assigned for endoscopic drainage. Several case series have described success rates ranging between 80 and 100% when FC-SEMS were used for the endoscopic drainage of WON [15‐18]. However, high rates of stent migration were also reported (15%) [19].

Lumen-apposing metal stents (LAMS) were designed in order to decrease the risk of stent migration. The stents have a saddle-shaped design with anchoring flanges and were originally applied to prevent leakage between the walls of both collections and the stomach (Fig. 1). Its large diameter also permits the spontaneous passage of necrotic tissue into the stomach, which may reduce the need for necrosectomy. LAMS also provide, if necessary, an easy entry port for endoscopic necrosectomy.

Over 30 retrospective case series, four prospective cohort studies and one interim analysis of a RCT evaluated the efficacy of LAMS for drainage of PFCs (Table 2). A recent review evaluated the endoscopic drainage of WON using DPS versus metal stents from published studies between 1990 and 2016 [20••]. The authors included 41studies, mainly retrospective and non-comparative studies, regarding the use of DPS and metal stents (total of 2213 patients). Only five of the studies included in the meta-analysis compared DPS directly to metal stents (note that the authors excluded four studies due to low total number of patients and disproportion of DPS versus metal stents) [21••, 22•, 23‐25]. Overall, resolution of WON was more likely to occur when using LAMS (91.5%) compared to DPS (80.9%) (OR 2.5, 95% CI 1.4–4.3, P = 0.001). No difference was found in the resolution of WON after the initial endoscopic procedure with LAMS (52.3%) compared to DPS (43.4%) (OR 1.4, 95% CI 0.56–3.6, P = 0.4).

To date, there are conflicting literature reports with respect to the safety of LAMS in WON drainage. The aforementioned meta-analytical review indicated that the use of LAMS (in the non-comparable studies) was associated with lower bleeding (6.2 vs. 12.6%, P = 0.007) and lower stent occlusion (7.5 versus 17.4%, P = 0.015) [20••]. However, the results of an ongoing randomized controlled trial about endoscopic drainage of WON with LAMS versus DPS have raised concerns [21••]. Stent-related adverse events were observed in 6 of the 12 patients randomized to LAMS: major bleeding (N = 3), buried stent syndrome (N = 2), and biliary stricture (N = 1) were reported [21••]. Interestingly, all adverse events occurred within 6 weeks after stent placement. As a result, the authors changed the study protocol by performing imaging 3 weeks after stent placement in order to remove the LAMS as soon as the necrotic collection was collapsed. These findings suggest that the risk of stent-related adverse events may be related to the timing of LAMS removal. Yet, current literature provides no consensus about the appropriate timing of LAMS retrieval: an interval ranging from 3.5 to 10 weeks after the initial drainage procedure has been described [21••, 26‐28]. Currently, the ESGE guidelines recommend to remove the LAMS after a maximum of 4 weeks [29••]. In order to minimize the risks of bleeding and migration, it has also been suggested to place an additional DPS through the lumen of the LAMS. A single-center retrospective was conducted to compare LAMS alone (N = 21) versus the combination of LAMS and DPS (N = 20) and showed that adverse events were significantly higher in the LAMS group compared to the LAMS combined with DPS group (43 versus 10%, P = 0.04) [30].

Endoscopic drainage with LAMS may be restricted by its high costs. To date, high-quality methodologic studies comparing the cost-effectiveness of the different stents used for endoscopic drainage are lacking. In a recent study from the USA, the costs of endoscopic drainage with LAMS per patient were higher compared to DPS ($20,029 vs. $15,941) [31]. However, more patients achieved successful endoscopic drainage of WON with LAMS (92 vs. 84%). For this reason, the authors described that the incremental cost-effectiveness ratio favored drainage with LAMS.

In contrast, the theoretical benefit of LAMS is not present in the endoscopic drainage of pancreatic pseudocysts. A meta-analysis showed no difference in decrease in PFC size and/or resolution of symptoms between patients with pseudocysts treated with DPS (85%, 95% CI 81–89%) or with metal stents (83%, 95% CI 74–89%) [32]. Although randomized trials are lacking, DPS should be preferred over LAMS when draining pseudocysts given their excellent safety profile and the possibility to leave DPS in for longer periods of time.

To summarize, endoscopic drainage with LAMS seems to result in higher resolution rates of WON when compared with DPS. Therefore, LAMS should be considered in patients with infected WON who are eligible for endoscopic drainage. Nonetheless, the best available evidence originates from only a few prospective studies. Moreover, important questions regarding the safety of LAMS are still unanswered. The current recommendation to remove LAMS after a maximum of 4 weeks is also not based on high-quality evidence [29••]. The suggestion that the placement of DPS within LAMS would minimize the risk of adverse events should be further looked into, although it theoretically conflicts with the proposed drainage advantages associated with the large diameter of LAMS. If a necrotic collection is not fully collapsed, LAMS should be exchanged for DPS, but this may not be easy or even possible in all patients. Finally, it is unclear whether LAMS are cost-effective. In conclusion, these concerns need to be addressed by future multicenter prospective studies before LAMS can be advised for routine use.

We have recently initiated a prospective multicenter clinical study investigating the use of LAMS (Hot AXIOS™) in the management of infected WON (AXIOMA study, NTR7056). With this study, we aim to evaluate the treatment outcomes and safety of LAMS. Furthermore, AXIOMA will test the hypothesis that the use of LAMS is more cost-effective by reducing the need of additional necrosectomies.

Besides the administration of large diameter metal stents, optimization of endoscopic drainage may be achieved by the creation of multiple transluminal tracts into the WON cavity, also called the “multiple transluminal gateway technique” (MTGT).

Three retrospective case series compared MTGT to conventional drainage techniques for the treatment of WON [33‐35]. In total, 39 of 204 patients (19%) were treated with MTGT. MTGT was applied in patients who responded inadequately to the initial endoscopic drainage procedure and in case WON exceeded ≥ 12 cm [33, 35]. A maximum of three transluminal tracts were created. Treatment success between 92 and 100% was reported in patients receiving MTGT compared to 52–70% receiving conventional drainage techniques. One disadvantage of MTGT was a significantly longer median procedural duration in comparison with the conventional drainage technique (37 vs. 22 min, P = 0.017) [35]. However, the introduction of a LAMS allowed a more rapid creation of multiple transluminal tracts (< 10 min) [36].

Transluminal drainage can be combined with percutaneous drainage in particular cases of WON extending to the paracolic gutters. This is also referred to as “dual-modality drainage” (DMD).

Five retrospective studies, which were all initiated in the same hospital, reported on the DMD technique [37‐41]. One of the studies was comparative and described 49 patients that completed DMD versus 45 patients that were only drained percutaneously [37]. It showed that the DMD technique was associated with a shorter length of hospitalization compared to percutaneously draining only (mean 24 vs. 54 days, P < 0.002). Additionally, DMD-treated patients required less drains (1.3 vs. 1.9, P < 0.001). Lastly, a lower number of endoscopic procedures (1.9 vs. 2.7, P < 0.02) and CT scans (7.8 vs. 14.0, P < 0.001) were needed. In another retrospective study, a total of 117 patients underwent DMD for either infected WON (N = 55) or symptomatic sterile WON (N = 62) [38]. Interestingly, none of the 103 patients that completed treatment (88%) needed an additional (surgical) necrosectomy. Furthermore, not a single patient developed a pancreaticocutaneous fistula after treatment with the DMD technique.

In conclusion, evidence suggests that MTGT and DMD are potentially effective approaches for the optimization of endoscopic drainage of WON. Although, these findings were obtained in retrospective studies only and further prospective studies are needed. Therefore, MTGT should only be considered in selected cases of patients who not respond well on the initial endoscopic drainage procedure or in the case of in large WON. Additional percutaneous drainage can be initiated if WON extends to the paracolic gutters.

One of the reported innovations is the use of hydrogen peroxide (H₂O₂) for the irrigation of WON. H₂O₂ decomposes into water and oxygen when combining with organic tissue and is therefore thought to facilitate the removal of necrotic debris.

Four retrospective studies and one prospective open label study reported on the use of H₂O₂ for the debridement of WON [42‐46]. A total of 108 patients underwent endoscopic drainage of WON followed by H₂O₂ irrigation. Drainage was performed with DPS, FC-SEMS, and LAMS. The reported number of required necrosectomy procedures ranged from 1 to 3. Resolution of WON occurred in 79 to 100% of patients.

Based on this limited set of data, there is currently no supportive evidence to recommend H₂O₂ in the treatment of WON. Moreover, its safety is questionable because several case series reported that its use may be related to embolic events [47‐52]. In conclusion, irrigation with H₂O₂ is currently not advised.

Irrigation of WON can also be performed by the placement of a nasocystic tube. The tube can be inserted next to DPS or through a deployed metal stent.

To date, prospective studies assessing the duration, type, or volume of irrigation with a nasocystic tube are lacking. In a retrospective comparative study, patients treated with a nasocystic tube reported a decrease in symptoms after 1 month in combination with a minimum of 30% decrease in WON size on imaging (85 vs. 63%, P = 0.03). Furthermore, the addition of a nasocystic tube to placed DPS decreased stent occlusion rates as compared to using the stent alone (13 vs. 33%, P = 0.03) [53].

It is not clear whether irrigation with a nasocystic tube offers any advantages when combined with a LAMS. In a large multicenter retrospective study, 22 of 68 patients with WON were drained via a LAMS with the addition of a nasocystic tube. In this study, there was not a difference in WON resolution in patients drained with or without nasocystic tube (90.9 vs. 95.6%, P = 0.59) [54].

Although data on the irrigation with a nasocystic tube is scarce, its use can be recommended when a significant amount of necrosis is present in the necrotic collection.

The role of gastric acid in the chemical debridement of necrotic tissue has been investigated as well. It is thought that the low pH of gastric acid facilitates the liquefaction of necrosis and prevents bacterial overgrowth. Because proton-pump inhibitors (PPIs) cause a long-lasting reduction of stomach acid production, they could have negative effects on the resolution of WON.

So far, only two studies have focused on the role of PPI and WON resolution [55, 56]. A retrospective multicenter study investigating patients with WON showed that more necrosectomy procedures were required to achieve WON resolution in patients using PPIs (N = 136) in comparison with those (N = 136) who did not use PPIs (median number of procedures of 4.6 vs. 3.2, respectively) [56]. Another retrospective study of 71 patients (54 PPI users and 17 non PPI users) reported a similar trend (mean endoscopic procedures 2 vs. 1.4) [55].

Although this limited set of studies is insufficient to strongly recommend the discontinuation of PPIs in all endoscopically treated patients with WON, this can be considered in particular cases in which there is no strong indication to continue their use.

Vacuum-assisted closure (VAC) is an established treatment method for wounds. The negative pressure of a vacuum-sealed sponge results via several mechanisms into wound closure. Endoscopic vacuum-assisted closure (EVAC) is increasingly being conducted as a new method to repair upper gastrointestinal defects [57]. EVAC might be an effective addition to endoscopic necrosectomy if established endoscopic treatment options have failed.

However, only four case reports described the use of a transluminally placed vacuum-assisted closure system in the management of WON [58‐61]. Moreover, its potential benefits may be limited because the system should often be exchanged, resulting in an added number of endoscopic interventions. Consequently, it is unsure whether EVAC will offer any advantages for the management of WON. Future studies should determine the efficacy of EVAC and address safety endpoints such as the risk of a pancreatic or intestinal fistula.