While direct percutaneous endoscopic jejunostomy (DPEJ) is not yet widely practiced, advantages include longer durability compared to percutaneous endoscopic gastrostomy with jejunal extension (PEGJ) and less invasiveness than surgical methods with similar rates of complications. DPEJ placement does not require special equipment and is similar in technique to widely practiced percutaneous endoscopic gastrostomy (PEG), and thus should be more widely available. This article will review the current state of DPEJ, specifically addressing when and in whom it should be considered, the technical approaches to placement, and complications to avoid.
Direct percutaneous endoscopic jejunostomy (DPEJ) was first described by Shike in 1987 as a method establishing long term enteral feeding in cancer patients with previous gastric resection.1 DPEJ is an endoscopic procedure that places a percutaneous feeding tube directly into the jejunum similar to a percutaneous endoscopic gastrostomy (PEG) that places a feeding tube directly into the stomach. There are several approaches to long-term jejunal access, each with unique advantages and disadvantages (Table 1). Percutaneous endoscopic gastrostomy with jejunal extension (PEGJ) remains the most common endoscopic approach due to high reported initial success rates, but tube malfunction is common and requires frequent re-interventions. Percutaneous gastrojejunal feeding tubes are increasingly being placed fluoroscopically, which while not requiring sedation, does involve radiation exposure and may have similarly high tube malfunction rates. Direct percutaneous jejunostomy can also be performed through interventional radiology (IR-J), though there are only limited reports in the literature and it is not widely available.2 There are several surgical jejunostomy (SJ) methods including laparoscopic and needle-catheter techniques, which have success rates approaching 100%. However they require multiple skin incisions, general anesthesia, and if attempted laparoscopically may need conversion to open surgery.3 While DPEJ is not yet widely practiced, advantages include longer durability compared to PEGJ and less invasiveness than surgical methods with similar rates of complications. DPEJ placement does not require special equipment and is similar in technique to the widely practiced PEG, and thus should be more widely available. This article will review the current state of DPEJ, specifically addressing when and in whom it should be considered, the technical approaches to placement, and complications to avoid.
General Indications for Jejunal Feeding
General indications for jejunal feeding are the need to bypass the stomach due to gastric dysfunction or previous resection. Previous gastric resection can render the stomach dysfunctional both from impaired emptying and increased propensity to reflux. In addition, gastric resection may make the stomach difficult to access percutaneously. More specific indications for direct jejunal feeding include: gastric obstruction, gastroparesis, high risk of aspiration, previously failed gastric feedings, post-surgical or other anatomy that precludes gastric access, or need for jejunal decompression/venting. Gastrostomy is generally avoided in patients with anticipated esophagectomy due to concern that PEG may compromise the integrity of a future gastric conduit.4,5
Indications for DPEJ are summarized in Table 2. Maple et al. 6 reported the largest series of 307 patients undergoing DPEJ. In this report 28% of DPEJ were placed due to GI cancer, 21% for gastroparesis, 19% due to previous gastric surgery including partial or total gastrectomy and 13% due to high clinical risk of aspiration. The outcomes of DPEJ placement for specific clinical indications are summarized below.
Patients with recurrent aspiration pose a difficult challenge for clinicians and suffer high morbidity and mortality.7 Percutaneous gastrostomy has not been shown to decrease aspiration and aspiration pneumonia. In patients with a history of aspiration, recurrent aspiration after gastrostomy placement has been reported in 1162% of patients.8 Outcomes after PEGJ placement have also been disappointing with multiple studies unable to demonstrate decreased aspiration.9,10,11,12 The poor results of PEGJ with recurrent aspiration may be explained by several factors: frequent PEGJ tube malfunction with clogging and proximal migration into the duodenum or stomach and possibly increased duodenal-gastric reflux due to an extension tube crossing the pylorus.
A few studies have reported decreased aspiration with direct jejunal feeding using SJ13 and DPEJ.14,15 Panagiotakis et al.reported 11 patients who underwent DPEJ for recurrent aspiration.15 After DPEJ placement monthly aspiration events decreased significantly from 3.39 to 0.42. Another retrospective analysis compared 205 DPEJ and 58 PEGJ patients and reported aspiration in 18.1% of DPEJ vs. 31.0% of PEGJ patients though this difference was not statistically significant once adjusted for bed-bound status.16 In another report 16/18 patients with aspiration pneumonia who underwent surgical jejunostomy did not have recurrent aspiration.13 While the literature is limited, direct percutaneous jejunostomy feeding may reduce aspiration in high-risk patients.
Gastroparesis is a common indication for DPEJ placement.6 Good outcomes were recently reported in 14 patients undergoing DPEJ due to severe gastroparesis secondary to perioperative vagus nerve injury during lung transplantation.17 In these patients, DPEJ placement led to prolonged successful enteral feeding in 90% of patients. If PEG is required for gastric venting, a separate gastrostomy tube can be placed during the same procedure. Combined PEG/DPEJ has been shown to have decreased re-intervention rates and improved patient satisfaction compared to PEGJ alone.18
Up to 5% of patients with chronic pancreatitis will fail conservative treatment and are candidates for long term enteral feeding. In these patients, jejunal feeding provides nutritional support while minimally stimulating the exocrine pancreas. In one report, long term jejunal feeding in CP obtained by PEGJ or DPEJ decreased the percentage of patients with pain from 96% to 23% while increasing mean weight by 4.5kg over 6 months.19 A majority of the literature on jejunal feeding in pancreatitis has utilized PEGJ; however DPEJ has been reported successfully in patients with chronic pancreatitis as well.20,21 While more studies are clearly needed, this preliminary data suggests that DPEJ may be an effective means for providing nutritional support in these patients.
Critically Ill Patients
Percutaneous feeding tubes are not often required in the critical care setting due to shorter length of stay, however indications for percutaneous jejunal feeding are similar to those in non-critically ill patients. Successful DPEJ use has been reported in 17 mechanically ventilated ICU patients requiring long-term jejunal feeding for previous aspiration pneumonia (9), failing gastric feeds (4), post-operative anastomotic leak (3) and duodenal obstruction (1).22 All had successful DPEJ placement with one complication of colonic perforation. The remaining 16 tolerated jejunal feedings at goal rates with no episodes of aspiration and 13 were ultimately discharged from the hospital with continued jejunal feedings.
Comparison of PEGJ vs. DPEJ
When considering long-term jejunal access, the endoscopist must choose between PEGJ and DPEJ approaches. PEGJ has higher reported success rates,16,23 is generally considered less technically challenging, and is more widely taught in fellowship programs. However, with its longer and smaller caliber tube, PEGJs are prone to frequent clogging and tube migration. Fortunato et al. reported of 102 PEGJs in children, the mean number of jejunal tube replacements was 2.2 tubes per patient (range 1-14) with median tube functional duration of 39 days (range 2-274).24 Tube failure was most commonly due to displacement (31%), clogged tube (22%), and mechanical failure (19%). Similarly high PEGJ malfunction rates were reported in a study following 75 adults with PEGJ placement over 9 months, 53% of patients had tube dysfunction and 41% of patients requiring tube removal or exchange.25 A limited number of studies have directly compared PEGJ and DPEJ with results supporting longer durability and fewer malfunctions with DPEJ. Zopf et al. performed a retrospective analysis of 205 DPEJ and 58 PEGJ procedures and found that DPEJ functioned longer (272 days ± 414 vs. 130 days ± 223, p = 0.023) with less clogging (9.9% vs. 40%, p= 0.002) and tube migration (5.4% vs. 33.3%; p = 0.005).16 Fan et al. found that over 6 months follow-up, DPEJ was associated with signiﬁcantly longer feeding tube patency and less re-intervention (13.5% DPEJ vs. 55.9% PEGJ, p= 0.002).23 These results are consistent with a retrospective study reported in abstract only demonstrating longer durability of combined PEG plus DPEJ compared to PEGJ alone.18 Finally a randomized prospective trial also reported only in abstract form reported fewer re-interventions and higher patient satisfaction with DPEJ over PEGJ.20
Contraindications to DPEJ (Table 2) are similar to PEG and include uncorrected coagulopathy (with standard threshold INR < 1.5 and platelets > 50,000), proximal GI obstruction that would prevent passage of the endoscope to the jejunum, and inability to oppose the small bowel to the anterior abdominal wall. Management of antithrombotics should follow published guidelines. PEG is considered a high-risk procedure for bleeding by the ASGE (American Society of Gastrointestinal Endoscopy) guidelines and similarly so should DPEJ placement.26 Relative contraindications include severe obesity, peritoneal dialysis, ascites and neoplastic, inflammatory, or infiltrative diseases of the small bowel and/or anterior abdominal wall. In addition, other standard general contraindications to endoscopy apply as well.
As recommended by ASGE Practice Guidelines27 nasoenteric feeding is the preferred approach for patients expected to be able to resume oral feeding within 30 days. Therefore DPEJ should be considered in the management of patients with an indication for jejunal access expected to require enteral feeding for greater than 30 days.
If it is highly likely that jejunal feeding will be needed for > 30 days, it is not necessary to wait until then to place DPEJ. Alternatively, DPEJ placement is never urgently indicated and thus naso and oro-enteric feeding tubes can be used until clinical parameters are maximized for safe percutaneous placement or to determine if clinical improvement will eliminate the need for longer term jejunal feeding.
As an extension of ASGE guidelines for PEG, preprocedure laboratory evaluation is generally not required for DPEJ placement. Prophylactic antibiotics, typically cefazolin 1gm IV, are indicated unless the patient is already receiving other adequate antibiotic coverage. Anti-spasmodics such as glucagon and hyoscine can be used during the procedure to help reduce jejunal peristalsis. The procedure is performed with the patient in the supine position. DPEJ may be performed under conscious sedation, monitored anesthesia care, or general anesthesia. Procedure times are greater compared to PEG and therefore at our institution we now perform all initial DPEJ placements under general anesthesia to decrease aspiration risk during a prolonged procedure in the supine position. Use of CO2 for insufflation is also useful to reduce postoperative discomfort and risk of bowel over insufflation. Standard commercially available PEG kits can be used for the procedure with the only caveat that the looped guide wire and snare are long enough to pass through whatever enteroscope is used for the procedure. The authors most commonly use a pediatric colonoscope for DPEJ placement, but use of dedicated balloon and non-balloon enteroscopes have been reported as well.
Technique (see Figure 1)
DPEJ is technically similar to a pull PEG procedure with the main difference that additional procedural time is required to reach the small bowel and identify an appropriate area for direct jejunal puncture. The enteroscope is advanced into the jejunum and an appropriate puncture site is identified by observing transillumination of light in a discrete area on the anterior abdominal wall. To further confirm the site, external pressure using a single finger should produce a discrete indentation in the jejunal wall visualized by the endoscope. Both excellent transillumination and discrete ﬁnger indentation are required for safe jejunal puncture. After sterile skin preparation and local anesthesia, a sounding needle (typically 21 gauge 1.5-inch needle) is advanced into the jejunal lumen using the “safe-track” technique (see Table 4). Once visualized in the jejunum (Image 1), the tip of the sounding needle is immediately secured with a snare to prevent migration of the jejunum away from the abdominal wall. Using a scalpel, a stab incision is made immediately adjacent to the sounding needle and the larger trochar/needle cannula (~14 gauge) is inserted into the jejunum parallel to the sounding needle again using safe track technique and the snare is transferred from sounding needle to the trocar. From this point the procedure follows the same steps as a standard pull PEG technique. The looped guidewire is fed through the trochar, snared and withdrawn with the endoscope through the mouth. The jejunostomy tube is attached to the wire and pulled into place until the bumper is seated in the jejunum (Image 2). Endoscopic re-visualization can be performed at the operator’s discretion to conﬁrm placement, but is not necessary in uncomplicated procedures.
Leaving the jejunal port open immediately after successful placement facilitates small bowel decompression. Tube feeding can be initiated immediately at 25-50mL/hour and titrated to goal rate as tolerated. Feeding with standard polymeric formulas can be used in almost all patients unless there is concern for malabsorption or pancreatitis. There is little information regarding the time for secure adherence of the jejunum to the abdominal wall, thus it is advisable to wait at least 4 weeks prior to removal of a DPEJ. It is likely that mature stoma tract formation occurs much earlier unless the patient has poor wound healing from steroids, diabetes mellitus, immunosuppression, or severe malnutrition.
In the larger studies DPEJ success rates have ranged from 65-86% (Table 3).6,16,28,30 Increased success is consistently reported in patients with prior gastric or esophageal resection. This is likely secondary to postoperative anatomy being conducive to passing the endoscope further into the small bowel and adhesions ﬁxing the jejunum to the anterior abdominal wall facilitating site identiﬁcation. In the largest series of DPEJ placement the most common reasons for failure were absent transillumination/finger indentation (20%), suboptimal transillumination/finger-indentation (5.5%), inability to reach the jejunum (2.6%), inability to pass the scope concomitant with failure to transilluminate (1.9%), and sedation complication (1.3%).6
As with other endoscopic procedures, DPEJ placement success appears to follow a learning curve. One study reported success rates of inexperienced operators (<66 procedures) was 63.7% compared to 77.5% and 74.5% in medium (67-187 procedures) and highly experienced endoscopists (>188 procedures) respectively.16 Obesity has also been shown to decrease successful DPEJ placement in two studies. The first reported patients with abdominal wall thickness on CT scan <3cm (vs. ≥3cm) had successful placement 73% vs. 39% (p= 0.003) of the time with no successful placements when thickness exceeded 4 cm.29 In the second study, increasing BMI was directly associated with decreased success rates with BMI < 18.5 (96% success), BMI 18.5-24.9 (81% success), BMI 25-29.9 (73% success) and BMI > 30 (60% success).30
If DPEJ is initially unsuccessful, PEGJ can be placed during the same procedure, although there is data that repeat DPEJ attempt under general anesthesia with a dedicated balloon enteroscope is often successful.31 A number of helpful hints have been described to aid in successful DPEJ placement (Table 5).
Several adjunctive methods to the standard DPEJ technique described above have been reported. Adding fluoroscopy to the procedure allows confirmation that the endoscope has passed the ligament of Treitz and can be a useful tool to visually guide the sounding needle to the snare.32 A success rate of 97.5% was reported in a series of 40 patients utilizing the selective use of a 15 cm long drainage needle with intermittent fluoroscopic guidance.33 The authors felt that using a longer needle facilitated deeper penetration in obese patients in whom traditional transillumination or jejunal indentation may not be feasible. The use of ultrasound during DPEJ has also been described to help delineate a targeted loop of jejunum, confirm close proximity to the anterior abdominal wall, and rule out intervening viscus when transillumination has failed. 34 Routine use of fluoroscopy and ultrasound have not been widely adopted, but they remain useful modalities to consider in potentially difficult patients.
Perhaps the most promising adjuvant technique for DPEJ placement has been the use of single35 and double-balloon enteroscopy (DBE). Song et al. reported a series of 10 patients who failed standard DPEJ with pediatric colonoscopes and subsequently underwent repeat DPEJ with DBE with 100% success.31 The reasons for initial failures were inability to transilluminate (5 patients), altered anatomy of efferent limb (3 patients), and limited endoscopic advancement due to looping and small-bowel angulation (1 patient each). The authors further reported 93% success (27/29 patients) utilizing DBE for initial DPEJ placements. In this series repeat DPEJ with DBE was performed under general anesthesia in almost all patients. The authors hypothesized that improved outcomes with DBE were due to the ability to pass further into the small bowel from reduced instrument looping and a more flexible tip on the DBE scope allowing passage through small bowel angulation and narrowing.
WHAT TO AVOID
DPEJ placement should be considered only with extreme caution for patients with small bowel motility disorders or for patients with eating disorders such as anorexia or bulimia, functional nausea or vomiting, or patients with chronic pain (Table 2). If there is concern that small bowel feeding would not be tolerated, a trial of nasojejunal feeding prior to more invasive DPEJ placement is warranted.
Reported complications from the larger DPEJ studies are summarized in Table 3. Complications are generally rare, but severe adverse events reported in 2-6.3% of cases. These include jejunal volvulus, necrotizing fasciitis, bowel perforation, and bleeding. Small bowel volvulus is a potential complication of jejunostomy regardless of placement technique as the relatively unfixed small bowel can rotate around the single point fixation of the jejunostomy. Small bowel volvulus is likely less common in patients with previous abdominal surgery due to adhesions that may fix the small bowel at multiple locations. Use of T-fasteners to establish additional points of jejunal fixation near the site of the DPEJ may reduce the risk of jejunal volvulus. As noted previously obesity decreases success rates and it has 34 also been associated with a trend towards more adverse events. In one series the adverse event rate was 19% in patients with BMI > 25 and 4 of 5 severe adverse events also occurred in patients with BMI > 25.30
DPEJ provides direct jejunal access without the morbidities of surgery. It should be considered whenever long-term jejunal access is required. The most common indications are altered gastric anatomy, gastroparesis, GI malignancies, and recurrent aspiration. Limited data suggests that DPEJ provides more reliable long-term functionality, fewer re-interventions, and decreased risk for aspiration compared to the more commonly practiced PEGJ. DPEJ use has been limited due to
perceived need for greater technical expertise, however
the procedure is very similar to the pull PEG technique with the distinction that greater time and caution is required while identifying the site for direct jejunal access. Placement success can be improved by using a number of helpful hints and adjuvant techniques including use of balloon enteroscopy. Complication rates are similar to other methods of direct jejunal feeding tubes and include volvulus, necrotizing fasciitis
and persistent jejunal fistula. At this time improvements
in DPEJ techniques and further prospective studies comparing DPEJ to other methods of jejunal access are needed, but present data suggest that DPEJ should be more frequently practiced.