Percutaneous Endoscopic Gastrostomy for Decompression of Nonmalignant Gastrointestinal Disease

Introduction

The percutaneous endoscopic gastrostomy (PEG) procedure was initially developed for establishing long-term enteral feeding access in those unable to ingest food orally. ¹ However, its indications eventually expanded to provide an alternate means for prolonged gastrointestinal (GI) drainage other than nasogastric (NG) tube decompression, which is difficult for patients to tolerate and can be fraught with complications.^2–5 The utility of PEG for decompression of inoperable malignant GI obstruction has been well documented, especially in those suffering from gynecological malignancies, such as ovarian cancer.^6–10 However, there are limited data on the use of decompressive PEG for nonmalignant indications like prolonged ileus, gut dysmotility disorders, or complicated nonmalignant bowel obstruction. Many of these patients suffer from ongoing obstructive symptoms while surgical intervention is deferred because of prohibitive medical or surgical issues. PEG insertion offers a potential bridging solution while nonoperative measures are used or complex procedures are planned. Furthermore, in carefully selected patients, it may also facilitate discharge of patients who would otherwise linger in the hospital with ongoing obstructive symptoms.

To date, only a few publications have addressed the use of PEG decompression for nonmalignant indications, and the small case numbers that have been published tend to only address feasibility as subset analyses within malignant decompression studies.^2,3,11–13 The aim of this study was to assess the safety and short-term outcomes of decompressive PEG for prolonged ileus, gut dysmotility, and complicated nonmalignant obstructions in the largest series presented to date.

Subjects and Methods

After approval by the Institutional Review Board, we retrospectively identified and reviewed all PEG insertions completed at our institute between 2009 and 2014. Cases were identified by our financial department based on billing and Current Procedural Terminology (CPT) codes, which also captured failed insertion attempts. We then excluded all cases that were performed for enteral or malignancy-related indications.

Patient demographics, procedural data, and postintervention parameters were then extracted for analysis from our computerized patient charting system. Patient demographics included age, sex, body mass index, comorbid conditions, number and type of previous abdominal surgeries, procedural indication or related diagnosis, and length of stay before PEG insertion. Procedural parameters included endoscopic operator, inpatient or outpatient location, findings, and complications. Postprocedural data extracted included postinsertion compilations, length of stay, and need for re-admission and/or re-operation up to 60 days after discharge.

Categorical variables were reported as frequencies (%). Continuous variables with normal and non-normal distributions were presented as mean ± standard deviation and median (range), respectively. Statistical analysis was performed in SPSS for Windows software (SPSS, Inc., Chicago, IL).

Results

In total, we identified 128 decompressive PEG insertions during the study period (malignant obstruction, n = 55). From those, we identified 72 patients (42 females, 30 males) with a mean age of 58.8 ± 15.2 years who had a decompressive PEG for nonmalignant indications (Table 1). The cohort had a mean body mass index of 24.2 ± 5.2 kg/m², with a median of four comorbidities (range, 1–14), and all but 1 patient had had previous abdominal surgery. Five (6.9%) cases were done as outpatient procedures, and all remaining PEGs were done during inpatient admissions.

Specific procedural indications are summarized in Table 1. There were no procedural complications or 60-day mortalities. PEG placement and NG tube removal were successful in all 72 cases. The basis of removal of all NG tubes prior to discharge and replacing them with PEG tubes was due to the issues with patient intolerance and associated nasopharyngeal irritation.

One (1.4%) outpatient PEG-associated bleed requiring a brief admission for nonoperative management was the only procedure-related complication. One case required emergency laparotomy because of concern for ongoing small bowel obstruction with clinical signs of ischemia 3 days after the PEG insertion. This patient had high operative risk (previous multi-organ transplant, dialysis-dependent end-stage renal disease) and was found to have a noncompromised internal hernia in a frozen abdomen at emergency laparotomy (obstruction resolved after extensive lysis of adhesions).

The mean pre- and postprocedural lengths of stay were 14.4 ± 10.7 and 7.6 ± 11.1 days, respectively (P = .0003). The 30-day re-admission rate was 12.5% (9 patients). Of these, five were for unrelated issues, two were brief observation admissions (both 3 days) for emesis with an appropriately functioning PEG, one was for definitive operative management of a chronic obstruction in a high-risk patient, and one was for a new enterocutaneous fistula (non–PEG-related). When expanding to 60 days after discharge, we found three more re-admissions: one for nausea and bloating (resolved without intervention), one for respiratory failure in a scleroderma patient, and one for recurrent adhesive small bowel obstruction necessitating laparotomy and small bowel resection. This last patient had a decompressive PEG placed for prolonged postoperative ileus after small bowel volvulus and subsequent resection (required a “second look” laparotomy for ileocolic anastomosis). He returned to the hospital 45 days after discharge with recurrent obstructive symptoms and was found to have a strangulated segment of bowel from an adhesive band.

Fifty-two (72%) of the 72 patients were discharged with a PEG for decompression for a median of 69.5 (range, 17–316) days; the remaining 20 (28%) patients were lost to follow-up and were assumed to continue follow-up with their primary referring center. Of the 72 patients, 63 (87.5%) were discharged on total parenteral nutrition (TPN); 36 (50%) of those continued to receive TPN for a median of 51 days (range, 4–316 days) after discharge, resolution of their mechanical obstruction was ultimately achieved, and they subsequently resumed enteral nutrition. Twenty (28%) patients were lost to follow-up; because they were referred to our tertiary referral center we assumed they continued follow-up at their referring institution. Three (4%) patients ultimately went on to use TPN indefinitely and thus were considered to be chronically TPN-dependent.

Discussion

Gauderer et al. ¹ first described the PEG procedure in 1980 as an effective means for obtaining long-term enteral feeding access, and this remains the most common indication for this endoscopic-guided intervention. Over the years, PEG has been shown to be a safe and effective procedure that can be performed with the patient under conscious sedation in most circumstances. The indications eventually expanded to include GI decompression, especially in the setting of malignant mechanical bowel obstruction, and many groups have validated this approach.^2–10 However, there are limited data on the use of decompressive PEG in the setting of prolonged ileus, dysmotility disorders, and complicated nonmalignant bowel obstruction. Publications addressing this sparsely utilized indication are few in number and report on small case numbers in studies focusing on malignant indications.^2,3,11–13

In one of the earliest reports, Picus et al. ¹¹ published results on six PEG insertions for chronic intestinal obstruction (five for malignancy-related obstruction, one for nonmalignant obstruction). The nonmalignant patient (duodenal obstruction from chronic pancreatitis) tolerated the procedure well and was able to have his NG tube removed while undergoing nonoperative management.

Around the same time, Stellato et al. ² reported on successful PEG decompression of chronic intestinal obstruction in 9 patients. All insertions were successful, but they did report one transhepatic gastrostomy insertion, which caused no adverse event.

Herman et al. ³ reported on 7 cases of decompressive PEG for stasis and nonmalignant obstruction in a study that mainly focused on malignant decompression (n = 46). All seven insertions were successful and required a mean of 31 days of decompression, with all tubes subsequently being removed when the underlying causes resolved. However, it is unclear in this report if any of the complications encountered occurred in the malignant or nonmalignant cohort. ³

Felsher et al. ¹² published their experience with 13 decompressive PEGs for nonmalignant indications, which included five fistulas after surgery/radiation, four cases of ileus/obstruction after laparotomy, two Crohn's disease–related cases of fistula/obstruction, one gastroparesis, and one for pseudo-obstruction. PEG insertion was successful in 11 of 13 cases, with the 2 failures being attributed to inability to locate safe percutaneous access. Five PEGs were inserted as a bridge to definitive surgical intervention with a mean of 6 months until surgery (1 patient was still awaiting surgery). The remaining 6 cases were for prolonged ileus or obstruction, with 4 of them achieving resolution and 2 remaining PEG-dependent. Again, this study included both malignant and nonmalignant cases, and it is unclear if any of the encountered complications was in the nonmalignant cohort. ¹²

In the present study, we focused solely on those who required decompression for nonmalignant indications. As previously mentioned, the CPT codes utilized also captured failed attempts, which were seen in cases for malignant indications but not in the study cohort. PEG decompressions for malignant obstruction tend to be more difficult than nonmalignant procedures because of disease-specific factors like ascites and carcinomatosis, and this likely explains why our success rate is higher than what is reported in malignant decompression studies.

The various nonmalignant indications for PEG insertion in our cohort are summarized in Table 1. Although there were no complications at the time of endoscopic-guided PEG insertion, we did encounter one procedure-related complication (PEG-associated bleed), and 1 case ultimately required emergency laparotomy for ongoing small bowel obstruction with clinical signs of gut ischemia. The former patient responded to nonoperative management (blood transfusion and observation), and the latter patient with high operative risk was found to have a noncompromised closed-loop obstruction in the setting of a frozen abdomen; this patient did well after extensive lysis of adhesions and was discharged with a complete return of intestinal function.

Perhaps the most obvious benefit in the cohort was the ability to decompress patients sufficiently to facilitate discharge from the hospital. NG tube decompression is uncomfortable for the patient and can preclude discharge home or to interim facilities; PEG decompression offers a more comfortable and easier-to-manage alternative. In the present study, the cohort spent a mean of 14.4 ± 10.7 (range, 1–60) days in the hospital before PEG decompression compared with 7.6 ± 11.1 (range, 1–71) days after PEG insertion (P = .0003), and all inpatients were eventually discharged. Although there were numerous other confounding factors that contributed to length of both pre- and postprocedural hospital stay, there was an obvious trend of rapid symptom resolution and discharge after PEG decompression. For instance, 30 of 67 (44.8%) inpatient cases were discharged within 4 days after PEG insertion. Ultimately, there were 12 (16.7%) 60-day re-admissions, but most were for unrelated issues or nausea/emesis with a properly functioning gastrostomy that quickly resolved without invasive intervention. There was also one planned admission for definitive operative management after the bridging PEG was inserted, 1 patient was admitted with a new enterocutaneous fistula related to his underlying disease (recent colectomy for Clostridium difficile colitis with subsequent ileorectal anastomotic leak and prolonged ileus), and there was one recurrent adhesive small bowel obstruction requiring an emergency laparotomy and bowel resection.

There are limitations to this study. First, the analysis was completed retrospectively, and all cases occurred at a single center. Furthermore, the majority of patients were referred to our tertiary center for their primary procedure or underlying disease process, and while re-admissions and need for operative intention after PEG insertion was captured in our data, PEG removal data was limited due to loss of follow-up.

Conclusions

Nonmalignant intestinal obstruction can be difficult to manage, especially when patients have prohibitive operative risk, severe intraabdominal adhesions, gut dysmotility disorders, or early postoperative ileus. Decompressive PEG can safely and effectively decompress nonmalignant intestinal obstruction in complex patients, ultimately allowing for NG tube removal and discharge from the hospital. It can also act as a bridge to definitive repair of complex abdominal pathology.