Gastrojejunostomy Complications Frequent

Gastrojejnostomy (GJ) placement allows enteral feeds to bypass the stomach.  When a gastrostomy is already in place, GJ placement may allow patients to avoid surgery (eg. fundoplication).  Most practitioners would consider the risk of GJ placement to be low, but a recent report (J Moorse et al. JPS 2017; suggests that it is higher than expected.  The abstract and link are below.

Link: Gastrojejunostomy tube complications — A single center experience and systematic review



Gastrojejunostomy tubes (GJTs) enable enteral nutrition in infants/children with feeding intolerance. However, complications may be increased in small infants. We evaluated our single-institution GJT complication rate and systematically reviewed existing literature.


With REB approval, a retrospective single-institution analysis of GJT placements between 2009 and 2015 was performed. For the systematic review, MOOSE guidelines were followed.


At our institution, 48 children underwent 154/159 successful insertions primarily for gastroesophageal reflux (n = 27; 55%) and aspiration (n = 11; 23%). Median age at first GJT insertion was 2.2 years (0.2–18). Thirty-five (73%) had an index insertion when ≤10 kg. GJTs caused 2 perforations and 1 death. The systematic review assessed 48 articles representing 2726 procedures. Overall perforation rate was estimated as 2.1% (n = 36 studies, 23/1092, 95% CI: 1.0–3.2). Perforation rates in children <10 kg versus ≥10 kg were estimated as 3.1%/procedure (95% CI: 1.1%–5.0%) and 0.1%/procedure (95% CI: 0%–0.3%), respectively. The relative risk of perforation was 9.4 (95% CI: 2.8–31.3). Overall mortality was estimated as 0.9%/patient (n = 39 studies; 95% CI: 0.2–1.6%). Most perforations (19/23; 83%) occurred ≤30 days of attempted tube placement.


Gastrojejunostomy tubes are associated with significant complications and frequently require revision/replacement. Insertion in patients <10 kg is associated with increased perforation risk. Caution is warranted in this subgroup.

With regard to the methodology

  • ~90% of the procedures were performed by interventional radiology and the interventionist had a median of 6.6 years of experience
  • Most GJs were 16 French in width and most were either 15 cm or 22 cm in length

My take: This report highlights the significant risks associated with GJ placement, particularly in smaller patients (<10 kg).  Despite these risks, GJ placement is often the safest option.

Costa Maya, Mexico

Costa Maya, Mexico


Adverse Events Following Pediatric Endoscopy –Underestimated Previously

A recent study (RE Kramer, MR Narkewicz. JPGN 2016; 62: 828-33) report the frequency of adverse events that occurred within 72 hours in a prospective observational cohort of 9577 patients from a single center.

The authors characterized complications more precisely and identified a much higher rate of complications than what has previously been reported.  Key findings:

  • The overall adverse event rate was 2.6% with 1.7% of all cases requiring unanticipated medical care.
  • Absolute risk of bleeding was 0.11%, infection 0.07%, and perforation 0.1% (n=12).  In total, these standard measures of complications were 0.28%.
  • Advanced and therapeutic cases had much higher rates of adverse events. Perforations occurred after esophageal dilatation (5), esophageal food impaction (1), polypectomy (4), and primary GJ placement (2).
  • Adverse rate with ERCP was 11.54%
  • Adverse rate with PEG was 10.71%
  • Adverse rate with dilatation was 10.94%.  It is noted that a total of 319 dilatations were reviewed.  5 had perforations.
  • Adverse rate with polypectomy was 6.27%.  It is noted that a total of 128 polypectomies were reviewed.  4 had perforations.
  • The authors did not identify a significantly higher complication rate with trainee physicians.

As noted in a previous entry (see below), studies in adults have an estimated a perforation rate of 0.09% and serious complication rates (GI and non-GI complications) of 0.15% for upper endoscopy and of 0.2% for colonoscopy. In addition, a large pediatric study of endoscopies, found a perforation rate of 0.014% for EGDs and 0.028% for colonoscopies. Thus, this report identifies a higher rate (10-fold) of perforation (driven by therapeutic endoscopy) and a much higher rate of adverse events, including 2.08% in diagnostic EGD and 3.9% for diagnostic colonoscopy.  Furthermore, for diagnostic EGD and for diagnostic colonoscopy, grade 2 (needing ER or unanticipated physician evaluation) or higher adverse events occurred in 1.21% and 2.31% respectively.

My take: Using a broader (and more accurate) definition of complications after endoscopy, the authors have demonstrated a much higher rate of adverse events, particularly following dilatation, PEG, polypectomy, and ERCP.  This report indicates that our preop counseling needs to be modified to inform families that complications are not quite so rare.

Related blog post:  High Endoscopy Complication Rate After Intestinal …

Complication -Unrelated to endoscopy:

pontine myelinosis

High Endoscopy Complication Rate After Intestinal Transplantation

A recent study (J Yeh et al. JPGN 2015; 61: 636-40) indicated a high rate of endoscopy complications in pediatric patients who have undergone intestinal transplantation.

Key points:

  • Complications: In this single-center study with 1770 endoscopies (1014 sessions), the serious GI complication rate was 1.8% (32/1770).  The complications included 11 GI perforations, 13 GI bleeds, 6 GI hematomas, 1 gastric mucosa avulsion, and 1 distention from retained air. The authors’ database was not designed to capture cardiopulmonary complications.
  • In comparison, the authors note that adults without intestinal transplantation have an estimated a perforation rate of 0.09% and serious complication rates (GI and non-GI complications) of 0.15% for upper endoscopy and of 0.2% for colonoscopy. In addition, a large pediatric study of non-transplant patient endoscopies, found a perforation rate of 0.014% for EGDs and 0.028% for colonoscopies. Thus, the authors are reporting a perforation rate (11 of 1770) that is more than 20-fold higher than this pediatric study’s colonoscopy perforation rate.
  • Their techniques are well-described. For example, “for ileoscopies, 2 to 3 sites each consisting of 2 to 3 biopsies were also taken every 5 to 10 cm from the distal graft…typically surveyed up to 50 to 60 cm from the ostomy or ileocolonic anastomosis.”
  • The reasons for endoscopy were most frequently related to diarrhea/stool output in 35% and for surveillance in 32%.
  • The other interesting finding was that “of histology-proven rejections, 45% had normal-appearing endoscopies.”

The authors recommend that patients with intestinal transplantation should have endoscopy at a specialized center with teams who are intimately familiar with these children.

My take: I worry that the high complication rates reported at this center may indicate that individuals (perhaps in training) who are less familiar with the patient’s anatomy are performing many of these endoscopies.  I think individuals very familiar with the patient’s anatomy are best-suited to perform these endoscopies; this may limit some individuals at these specialized centers and may include some skilled endoscopists outside of intestinal transplant centers.

Related blog postSomething Bad is Going to Happen | gutsandgrowth




Four IBD Articles –Four Teaching Points

  1. Inflamm Bowel Dis 2014; 20: 1891-1901
  2. Inflamm Bowel Dis 2014; 20: 1996-2003
  3. Inflamm Bowel Dis 2014; 20: 2013-21.
  4. Inflamm Bowel Dis 2014; 20: 2056-66.

The first article is listed as a ‘basic science’ article.  However, it has direct relevance to the clinical problem of anti-TNF-induced psoriasiform skin lesions.  The article notes that this problem affects about 5% of patients treated with anti-TNF agents.  The authors found that IL-36γ and IL-17C are increased in anti-TNF-induced psoriasiform skin lesions of patients with Crohn’s disease (n=13 patients).  An important clinical point was that 7 of these patients with “severe anti-TNF induced skin lesions were successfully treated with the IL-12/IL-23 neutralizing antibody ustekinumab.”  This was superior to topical steroids or topical tacrolimus.

The second article is a proof-in-principle type article showing that proactive measurements of infliximab (IFX) levels may improve outcomes.  This retrospective observational study examined 48 patients who had proactive IFX levels. In 12 of 48, IFX dosing was escalated after the first proactive monitoring.  In addition, over the study period, 15% of patients had their dosing lowered. In those with proactive monitoring, the probability of remaining on IFX was >80%.  Those patients who achieved trough levels >5 mcg/mL had >90% probability of remaining on IFX therapy.  The authors hypothesize that better IFX levels may reduce anti-infliximab levels.

The third article examines carbohydrate intake in relation to the development of Crohn’s disease (CD) and ulcerative colitis (UC). The authors utilized the “EPIC” cohort (European Prospective Investigation into Cancer and Nutrition) (Public Health Nutr 2002; 5: 1113-24).  among 401,326 enrollees at recruitment, the dietary intakes of carbohydrates were measured using validated food frequency questionnaires. In this cohort, 110 developed CD and 244 developed UC during followup. Key finding: there was no significant risk for IBD based on total carbohydrate intake.  This study does not exclude the possibility that specific carbohydrates could have an etiological role.

The fourth article, a case-control study (2002-2011),  examines risk factors for endoscopy-associated perforation and perforation-associated complications (PAC) in patients with and without IBD. n=217,334 lower endoscopies (with 9518 in IBD patients).  Perforation rates: 18.91 per 10,000 and 2.50 per 10,000 for IBD and non-IBD endoscopy respectively.  The use os systemic corticosteroids at the time of endoscopy was associated with a 13 times greater risk for PAC.

Related blog posts:

Magnets: Clear and Present Danger

As noted in previous blogs (see below for links), ingested magnets represent a significant problem.  Two more studies add data to this issue:

  • JPGN 2013; 57: 14-17.
  • JPGN 2013; 57: 18-22.

The first study relates a retrospective single-hospital experience from 1995-2012 (from Boston).  In total, 112 cases of magnet injuries were identified using an initial computer search followed by manual chart review.  The mean patient age was 6 years.  The incidence rate ratio of 3.44 during the period 2007-2012 indicates a significant uptick in these injuries compared with the prior period. In addition, “office toys” accounted for 18 of 45 injuries during this latter period. Nature of injuries: swallowed magnets accounted for 86% of these injuries with 13 (12%) requiring endoscopic removal and 4 (4%) needing surgical intervention.

The second study analyzed a nationally representative same from the US Consumer Product Safety Commision Database (NEISS) for emergency department (ED) visits involving magnet ingestion in children (< 18 years) from 2002-2011.  NEISS sample includes >100 hospitals and 7 children’s hospitals.  The authors note that searching NEISS is not always straight-forward as specific product codes need to be entered like toy, kitchen gadget, and others.  The findings:

  • 16,386 suspected magnet ingestion related ED visits from 2002-2011
  • 59.4% were boys , 54.7% were younger than 5 years
  • There was a >8-fold increase in visits for these ingestions from 2002 to 2011.
  • Study limitations: NEISS database –likely underestimates problem, and no specific NEISS code for magnets

The authors note that one institution has started screening for magnets prior to MRI after a 5 year-old with unrecognized magnet ingestion developed intestinal perforations after undergoing an MRI for torticollis.

Take-home message: Magnets are, in the words of Tom Clancy, a ‘Clear and Present Danger’ as ingestion of more than one magnet creates a high risk of perforation or fistula.

Related blog posts: