Percutaneous endoscopically assisted transenteric full-thickness gastric biopsy: initial experience in humans

ORIGINAL ARTICLE: Clinical Endoscopy

Percutaneous endoscopically assisted transenteric full-thickness gastric biopsy: initial experience in humans Christopher N. Andrews, MD, MSc, FRCPC, Paul Mintchev, Emil Neshev, MD, Hughie F. Fraser, MD, FRCPC, Martin Storr, MD, Oliver F. Bathe, MD, FRCSC, Stefan J. Urbanski, MD Calgary, Alberta, Canada

Background: GI neuromuscular diseases (GINMD) can cause severe dysmotility and symptoms. Full-thickness biopsy specimens may help diagnose these disorders histologically. Objective: To assess a novel percutaneous endoscopically assisted transenteric (PEATE) biopsy method for obtaining full-thickness gastric tissue in patients with suspected GINMD. Design: Prospective proof-of-concept case series. Setting: Tertiary care gastroenterology unit. Patients: Ten patients (8 women, mean [standard deviation] age 43 [10] years) with gastroparesis-like symptoms (mean [standard deviation] gastroparesis cardinal symptom index 3.28 [1.46] out of 5) and/or clinical findings suggestive of a gastric GINMD. Interventions: All patients underwent PEATE biopsy during standard gastroscopy as an outpatient procedure. Tissue was stained for histology and immunohistochemistry of gut wall elements. Interstitial cells of Cajal (ICC) counts were compared with archived normal gastric tissue from control gastrectomies. Main Outcome Measurements: Biopsy success, complications, histopathological findings according to the London Classification of GINMD. Results: Full-thickness antral tissue suitable for analysis was obtained in 9 in 10 patients (90%). PEATE biopsy was well tolerated by all patients without complications. Histology suggested GINMD in 4 of 9 cases (44%), with possible degenerative leiomyopathy in 2, probable inflammatory leiomyopathy in 1, and abnormal ICC networks (⬎50% reduction in ICC counts) in 1 patient. Limitations: PEATE biopsy specimen size is smaller than a standard laparoscopic full-thickness biopsy. Conclusions: PEATE full-thickness gastric biopsy is a simple and safe method of assessing histopathological abnormalities in gastric GINMD without the need for laparoscopy or general anesthesia. ( Gastrointest Endosc 2011;73:949-54.)

Abbreviations: DGP, diabetic gastroparesis; ENS, enteric nervous system; GINMD, GI neuromuscular disease; GP, gastroparesis; ICC, interstitial cell of Cajal; IGP, idiopathic gastroparesis; PEATE, percutaneous endoscopically assisted transenteric approach; SDAP, severe dyspepsia with abdominal pain. DISCLOSURE: Dr Andrews has filed a provisional patent for the PEATE procedure in the United States. The other authors disclosed no financial relationships relevant to this publication. Dr Andrews was funded in part by Calgary Laboratory Services R&D grant RE7133. Copyright © 2011 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 doi:10.1016/j.gie.2010.12.037

Current affiliations: Centre for Digestive Motility (C.N.A., P.M., E.N., H.F.F., M.S.), Division of Gastroenterology, Department of Medicine, Departments of Surgery and Oncology (O.F.B.) and Department of Anatomic Pathology (S.J.U.), University of Calgary, Calgary, Alberta, Canada. Presented at Digestive Disease Week, May 1-6, 2010, New Orleans, Louisiana (Gastrointest Endosc 2010;71:AB144). Reprint requests: Christopher N. Andrews, MD, MSc, FRCPC, Division of Gastroenterology, University of Calgary Medical Clinic, 3330 Hospital Drive NW, Calgary, AB T2N 4N1, Canada. If you would like to chat with an author of this article, you may contact Dr Andrews at [email protected].

Received November 1, 2010. Accepted December 28, 2010.

www.giejournal.org

Volume 73, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 949

Full-thickness gastric biopsy

GI neuromuscular diseases (GINMDs) are typically severe disorders of gut motility that are diagnosed based on symptoms, the absence of physical obstruction, and tests of function. Gastroparesis (GP), the classic GINMD of the stomach, has an increasing incidence and a huge morbidity and financial burden.1,2 Tests of function, such as gastric-emptying studies in GP or colonic transit studies in colonic inertia, are helpful for management but do not address the underlying etiology of the disorder. A recent advance in the diagnosis of GINMD has been made through analysis of full-thickness enteric tissue obtained via laparoscopy.3,4 Full-thickness tissue is required to examine the enteric nervous system (ENS) and circular and longitudinal smooth muscle fibers, which are not accessible with standard endoscopic mucosal biopsy forceps.5 Numerous case series have associated histopathological abnormalities with GINMD, and guidelines for histological techniques and reporting5 and a formal classification (the London Classification)6 of neuromuscular pathology have recently been published. In the stomach, recent evidence has shown histopathological disruption of the ENS in full-thickness stomach specimens from patients with GP.7-9 Obtaining gastric ENS tissue, to date, has typically required a laparoscopy or laparotomy. Experimental endoscopic approaches have either failed to access ENS tissue10 or have been associated with a high perforation rate.11 The need clearly exists for a simple and safe endoscopic method to obtain fullthickness gastric biopsy specimens, ideally without specialized equipment (such as required with natural orifice transluminal endoscopic surgery methods) and within the skill set of any competent endoscopist. We recently reported such a technique using a percutaneous endoscopically assisted transenteric (PEATE) approach in a dog model.12 Our aim was to perform this technique in patients with suspected GINMD and assess its safety and outcomes.

METHODS This study was approved by the University of Calgary (Conjoint) Health Research Ethics Board, and all participants gave written informed consent before undergoing the procedure. Participants were recruited for the study from the clinical practices of the investigators. Patients with symptoms and a clinical history suggestive of severe upper gut dysmotility underwent full diagnostic evaluation including gastroscopy, scintigraphic gastric-emptying testing, intra-abdominal imaging, and blood work to rule out possible systemic causes of dysmotility. Based on history and gastric-emptying test results, patients were categorized as having diabetic GP (DGP), idiopathic GP (IGP), or severe dyspepsia with abdominal pain (SDAP). The SDAP group had symptoms of GP but with gastric emptying not slow enough for a formal diagnosis of GP (defined as ⬎10% of a meal remaining in the stomach at 4 hours).13 950 GASTROINTESTINAL ENDOSCOPY Volume 73, No. 5 : 2011

Andrews et al

Take-home Message ●

A full-thickness gastric biopsy method such as percutaneous endoscopically assisted transenteric approach has the potential to allow a wide spectrum of patients with disorders of upper gut function to be tested for GI neuromuscular disease (GINMD) in a simple, safe, and minimally invasive manner, potentially improving our understanding and management of GINMD.

Patients with established diabetes mellitus were categorized as DGP even if their gastric-emptying test results did not show greater than 10% retention at 4 hours. Symptom severity in the 2 weeks before biopsy was characterized by using the Gastroparesis Cardinal Symptom Index, a validated measure of abdominal symptoms in this population.14 The PEATE technique was similar to the method previously described.12 Patients were put on a liquid diet for up to 3 days before the procedure (depending on the amount of retained food observed in the stomach at previous upper endoscopy procedures), fasted overnight, and were given a single dose of cefazolin 2 g intravenously before EGD. Fentanyl and midazolam conscious sedation was used in all patients except 1, who had propofol sedation under the supervision of an anesthesiologist. After a full endoscopic examination was performed, a suitable biopsy area was chosen based on indentation of the anterior stomach wall by external finger pressure on the abdominal skin and by transillumination with the endoscope. After cleansing of the abdominal skin and using a strict sterile technique, the skin and biopsy tract were anesthetized with lidocaine 2% using a 22-gauge tracking needle. The anesthetic needle was used to confirm a direct tract to the stomach, which was verified endoscopically. A 4-mm incision was then made through the abdominal skin, and a spring-loaded biopsy gun (Cook Quick-Core; Cook Medical Inc, Bloomington, Ind) with a 14-gauge coaxial needle (20-cm length, 20-mm throw length, 1.6-mm internal diameter) was inserted in its neutral configuration (Fig. 1). After penetrating the stomach under direct visualization, the needle was cocked, and the internal notched needle was extended. The notched needle was then withdrawn so that the notch straddled the stomach wall and then fired. The needle assembly was then removed and the tissue collected for processing. Up to 4 separate antral biopsy specimens were taken from each patient via the same abdominal wall incision. No closure procedure was routinely performed either intraluminally or on the skin. A small bandage was placed over the skin postoperatively. Patients were monitored for at least 3 hours before discharge. Biopsy specimens were immediately preserved in 10% neutral-buffered formalin and then paraffin embedded, www.giejournal.org

Andrews et al

Figure 1. PEATE biopsy instrument (Cook Quick-Core, 14 gauge). Image inset shows magnified view with internal needle extended and biopsy notch exposed.

sectioned, and mounted by using standard techniques. Biopsy specimens were routinely stained with hematoxylin and eosin and Masson’s trichrome, and immunohistochemistry for CD117 (c-kit) and PGP9.5. All samples were evaluated by a skilled GI pathologist (S.J.U.). The cellular processes of intramuscular interstitial cells of Cajal (ICC) per high-power field (⫻400) were counted over the entire muscle area of each biopsy specimen and expressed as a mean. Patient ICC values were compared with archived normal full-thickness gastric tissue from 13 control gastrectomies for stomach cancer. All control tissue was taken from visually and microscopically normal antrum and stained in a manner identical to that in the patient group. ICC numbers were defined as decreased if the average count was 50% or more less than the average count of the control group.5

Full-thickness gastric biopsy

clip remaining in situ. No further closure intervention was required after biopsy in any other patients. The procedure in most patients required 2 needle biopsy sets because repetitive passage of the cannula through the abdominal wall led to blunting and less reliable acquisition of tissue because of dulling of the cutting surface of the needle. Thus, full-thickness tissue was obtained in approximately 50% of needle passes, with the first firing of a needle set typically being the best. The mean duration of the entire procedure from endoscope insertion to removal was 27.8 minutes (range 21-35 minutes), which included time for a full endoscopic examination. Most patients experienced mild abdominal wall pain for up to 24 hours after the procedure. There was no suggestion of peritonitis, bleeding, or significant discomfort in any of the patients immediately or at 4-day and 30-day postoperative follow-up. Sufficient tissue suitable for histopathological analysis, defined as identifiable subserosa, circular and longitudinal muscle, and mucosa, was obtained in 9 of 10 patients (90%). One patient had no gastric tissue retrieved despite 4 passes of the PEATE needle. A representative normal PEATE biopsy specimen is shown in Figure 2. The mean (standard deviation) surface area of each mounted biopsy specimen was 8.5 (2.7) mm2, with a range of 18 to 38 high-power fields (400⫻) assessable per biopsy. Two patients had definite intramuscular fibrosis suggestive of a degenerative leiomyopathy, and 1 patient had findings consistent with an inflammatory leiomyopathy. One IGP patient met criteria for abnormal ICC networks, with decreased counts of ICC processes compared with control values. Histological findings are presented in Table 1, and ICC counts are shown in Figure 3. Overall, qualitative or quantitative abnormalities were seen in 4 of 9 of PEATE biopsy specimens (44%).

RESULTS DISCUSSION Ten patients (8 women, mean age 43 years, standard deviation ⫾10 years, range 33-66 years) were recruited. Four had DGP, 3 had IGP, and 3 had SDAP. All patients had symptoms for more than 1 year. One IGP patient had clinical evidence of a pan-gut motility disorder (intestinal pseudo-obstruction), and 1 SDAP patient had quiescent Crohn’s disease with repeatedly negative findings on imaging and endoscopy for any signs of active inflammation or structural obstruction, but with clinical gastric stasis symptoms and a massively distended stomach. Most patients were severely symptomatic, with a mean (standard deviation) Gastroparesis Cardinal Symptom Index score of 3.28 (1.46) out of 5. Demographic profiles are shown in Table 1. The procedure was well tolerated by all patients. The first patient had an endoclip placed because of mild postbiopsy oozing; postprocedure hemoglobin levels were unchanged. A follow-up endoscopy on this patient 121 days later showed no abnormalities except for the endowww.giejournal.org

This study provides the first description of a simple and safe endoscopic method to obtain full-thickness gastric tissue in humans. Although the number of cases in this initial series is small, there was no significant morbidity and no mortality from the PEATE full-thickness biopsy. Although laparoscopic methods also appear very safe, they do entail greater morbidity, with an open conversion rate of 2% and readmission rate of 8% in 1 large multicenter series.3 Laparoscopy also adds greater expense because of requirements for general anesthesia and an overnight stay. The PEATE technique is straightforward, similar in some aspects to other routine endoscopic procedures such as insertion of a PEG feeding tube. The procedure was performed on an outpatient basis by an active clinical gastroenterologist without advanced therapeutic endoscopy training. The ease of performing this technique represents an advance in the diagnosis of GINMD by allowing widespread ascertainment of these relatively rare disorVolume 73, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 951

Full-thickness gastric biopsy

Andrews et al

TABLE 1. Patient demographics and PEATE histological findings Patient

Age, y

Sex

GET*

1

44

F

Very delayed

2

42

F

Very delayed

3

66

F

Normal

4

55

M

Delayed

5

34

F

6

40

7

Comment

Diagnosis

Total GSCI

Histological diagnosis

IGP

3.67

Reduced ICC

DGP

3.58

Normal

SDAP

3.25

Normal

Diabetic

DGP

4.00

Normal

Very delayed

Diabetic

DGP

2.11

Unsuccessful

F

Delayed

Diabetic

DGP

1.44

Intramuscular fibrosis suggestive of degenerative leiomyopathy

45

F

Very delayed

IGP

3.86

Normal

8

33

F

Normal

SDAP

4.39

Normal

9

40

F

Delayed

Quiescent Crohn’s disease

SDAP

4.83

Intramuscular fibrosis suggestive of degenerative leiomyopathy

10

34

M

Very delayed

Intestinal pseudo-obstruction

IGP

0.17

Consistent with inflammatory leiomyopathy

Diabetic

PEATE, percutaneous endoscopically assisted transenteric approach; GET, gastric-emptying test; GCSI, Gastroparesis Cardinal Symptom Index (maximum score is 5); IGP, idiopathic gastroparesis; ICC, interstitial cells of Cajal; DGP, Diabetic gastroparesis; SDAP, severe dyspepsia and abdominal pain. *Gastric emptying defined as very delayed if ⬎10% retained at 4 hours.

Figure 2. Representative PEATE gastric biopsy specimen with normal histology showing layers of gastric wall. H&E, orig. mag. ⫻40. CM, circular muscle fibers; G, ganglion; LM, longitudinal muscle fibers; MUC, mucosa; SS, subserosa.

ders and thus accumulation of a larger normative patient group for analysis; however, larger studies are needed to further evaluate for safety and accuracy. The gastric biopsy specimens taken in this study were from the anterior aspect of the proximal antrum midway between the greater and lesser curvatures, the region suggested in the recent International Working Group guidelines on histological techniques for GINMD.5 An added advantage is that multiple passes in different locations will also reduce the likelihood that a patchy pathological process will be missed. The tissue collected is also sufficient for assess952 GASTROINTESTINAL ENDOSCOPY Volume 73, No. 5 : 2011

Figure 3. ICC counts of controls and patients. Mean number of intramuscular ICC per high-power field was calculated for each subject and plotted. Solid lines show group means. Dashed line shows threshold for diagnosis of abnormal ICC networks, defined as less than 50% of the control mean.

ment of all aspects of the gut wall and ENS, including staining for qualitative assessment (hematoxylin and eosin), neurons, muscle, ICC, and glia if desired from a single www.giejournal.org

Andrews et al

full-thickness biopsy. This is an important distinction because practices among laboratories performing histopathological assessments for GINMDs are highly variable.15 Diagnoses presented in this study were made according to the recently described London Classification,6 with allowances for the fact that the small tissue sample size may reduce diagnostic certainty. Further applications for the PEATE biopsy could also include serial biopsy in the same patient to assess changes over time or response to therapy. Because the study of treatment and management of GINMDs is in its infancy, there is no standardized approach at present to these histological abnormalities. Results of these biopsies are therefore predominantly of academic interest at this stage and may not change patient management or outcome. This study focused on describing the feasibility and safety of the PEATE technique, with diagnostic outcome data planned for publication subsequently. However, the biopsy results were helpful in establishing a parallel diagnosis in the patient with Crohn’s disease (patient 9), which can now focus on GINMD management as opposed to repeated and fruitless evaluations of the status of her quiescent Crohn’s disease. The patient with inflammatory leiomyopathy (patient 10) was referred for a laparoscopic full-thickness biopsy for confirmation, and empirical immunosuppression is planned. As suggested previously, a repeat PEATE biopsy will be performed for longitudinal assessment of tissue response after immunosuppression has been started. Although the sensitivity of the PEATE biopsy for the diagnosis of GINMD is currently unknown, patients without gross abnormalities on their biopsy specimen (ie, “normal” histology) may be reassured by this finding or may have management subsequently guided toward treatment of visceral hypersensitivity or investigation of other causes of their symptoms. The major disadvantage of the PEATE biopsy is the small specimen size compared with that obtainable by laparoscopy, in which a 1.5-cm2 section of stomach is suggested by the same guidelines.5 This allows ample tissue for archiving and research purposes, which is not possible with a PEATE biopsy specimen. Clearly, if certain histopathological changes are few or widespread, the sensitivity of PEATE will be less than that of a laparoscopic section. Although not directly comparable, the diagnostic yield for GINMD with full-thickness biopsy by laparoscopy was as high as 89% at the jejunum in 1 large series3 compared with 44% at the antrum in this study. The selection of a laparoscopic or PEATE method for fullthickness gastric biopsy will likely be a trade-off between accessibility and risk against sensitivity and certainty of diagnosis. This limitation could be offset by larger-gauge biopsy and other needle modifications. Alternatively, a PEATE biopsy could be considered as a routine first-line approach in the stomach, with subsequent laparoscopic tissue sampling if a major therapeutic intervention is conwww.giejournal.org

Full-thickness gastric biopsy

sidered based on results (eg, immunomodulation therapy). A second disadvantage of the PEATE method is that at present it is limited to biopsy of the stomach. Although biopsy of the small intestine is anticipated, biopsy of the colon with this method would need to be approached with caution because of infection risk. Obese patients and those with previous surgeries may make site selection for PEATE biopsy difficult (because of failure to transilluminate or impeding tracking anesthesia needle entry into stomach, for example). Once in the stomach, when the biopsy gun is pulled back so that the needle notch straddles the stomach, caution must be used to prevent pulling back too far. Otherwise the cannula may be pulled across the parietal peritoneum into the abdominal wall. If the gun is fired in this location, tissue collection will be impeded, with collection of abdominal skeletal muscle, fascia, or often no tissue at all. This is presumably what happened with the patient whose biopsy was unsuccessful. Peritonitis could be considered a potential risk of this procedure, but was not seen in this study and is generally considered to be a minimal risk in human patients because of the substantial thickness of the stomach relative to the biopsy size. The PEATE biopsy specimen is taken with the stomach maximally distended; as the stomach is deflated, the gastric muscle contracts, reducing the size of the defect and self-sealing rapidly. The omentum also likely inhibits leakage as it does for other sites in the gut. Mucosal clips may be applied endoscopically, if necessary, to reduce bleeding, although this was not performed after the first patient without any perceived increase in risk. Peritonitis could also potentially occur because of infection introduced by the percutaneous approach, but this was minimized by adhering to a strict sterile technique for the procedure and intravenous administration of antibiotic before the procedure, similar to guidelines for antibiotic prophylaxis for PEG tube insertion.16 The risk of bleeding is minimized by the fact that this biopsy is not done blindly. If significant bleeding were to occur, it would be expected as soon as the needle set was passed into the stomach; thus, action could be taken before the cannula was fired to minimize damage. Bleeding that is not endoscopically visible (eg, into the peritoneal cavity) is typically minor, based on experience from PEG tube insertion. Risks associated with gastric food retention, which are always substantial when dealing with GP patients, are minimized by having the patient fast, typically overnight, as well as having the patient adhere to a strict diet consisting only of clear fluids for up to 3 days (depending on the severity of the patient’s dysmotility) before the procedure. Assessment of gut wall histology appears promising in GP, and the recent publication of the London Classification of GINMD sets the framework for an exciting new avenue of GI research. From the clinician’s point of view, a tool that can help objectively differentiate organic causes Volume 73, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 953

Full-thickness gastric biopsy

of gut dysmotility from visceral hypersensitivity, for example, would be very welcome. The PEATE technique provides a simple and safe method to both understand and assess these disorders, which have substantial morbidity. REFERENCES 1. Jung HK, Choung RS, Locke GR 3rd, et al. The incidence, prevalence, and outcomes of patients with gastroparesis in Olmsted County, Minnesota, from 1996 to 2006. Gastroenterology 2009;136:1225-33. 2. Wang YR, Fisher RS, Parkman HP. Gastroparesis-related hospitalizations in the United States: trends, characteristics, and outcomes, 1995-2004. Am J Gastroenterol 2008;103:313-22. 3. Knowles CH, Veress B, Tornblom H, et al. Safety and diagnostic yield of laparoscopically assisted full-thickness bowel biospy. Neurogastroenterol Motil 2008;20:774-9. 4. Tornblom H, Lindberg G, Nyberg B, et al. Full-thickness biopsy of the jejunum reveals inflammation and enteric neuropathy in irritable bowel syndrome. Gastroenterology 2002;123:1972-9. 5. Knowles CH, De Giorgio R, Kapur RP, et al. Gastrointestinal neuromuscular pathology: guidelines for histological techniques and reporting on behalf of the Gastro 2009 International Working Group. Acta Neuropathol 2009;118:271-301. 6. Knowles CH, De Giorgio R, Kapur RP, et al. The London Classification of gastrointestinal neuromuscular pathology: report on behalf of the Gastro 2009 International Working Group. Gut 2010;59:882-7. 7. Forster J, Damjanov I, Lin Z, et al. Absence of the interstitial cells of Cajal in patients with gastroparesis and correlation with clinical findings. J Gastrointest Surg 2005;9:102-8.

Andrews et al

8. Harberson J, Thomas RM, Harbison SP, et al. Gastric neuromuscular pathology in gastroparesis: analysis of full-thickness antral biopsies. Dig Dis Sci 2010;55:359-70. 9. Iwasaki H, Kajimura M, Osawa S, et al. A deficiency of gastric interstitial cells of Cajal accompanied by decreased expression of neuronal nitric oxide synthase and substance P in patients with type 2 diabetes mellitus. J Gastroenterol 2006;41:1076-87. 10. Rajan E, Gostout CJ, Lurken MS, et al. Evaluation of endoscopic approaches for deep gastric-muscle-wall biopsies: what works? Gastrointest Endosc 2008;67:297-303. 11. Rajan E, Gostout CJ, Lurken MS, et al. Endoscopic “no hole” full-thickness biopsy of the stomach to detect myenteric ganglia. Gastrointest Endosc 2008;68:301-7. 12. Fraser H, Neshev E, Storr M, et al. A novel method of full-thickness gastric biopsy via a percutaneous, endoscopically assisted, transenteric approach. Gastrointest Endosc 2010;71:831-4. 13. Tougas G, Eaker EY, Abell TL, et al. Assessment of gastric emptying using a low fat meal: establishment of international control values. Am J Gastroenterol 2000;95:1456-62. 14. Revicki DA, Rentz AM, Dubois D, et al. Development and validation of a patient-assessed gastroparesis symptom severity measure: the Gastroparesis Cardinal Symptom Index. Aliment Pharmacol Ther 2003;18:141-50. 15. Martin JE, Hester TW, Aslam H, et al. Discordant practice and limited histopathological assessment in gastrointestinal neuromuscular disease. Gut 2009;58:1703-5. 16. Banerjee S, Shen B, Baron TH, et al. Antibiotic prophylaxis for GI endoscopy. Gastrointest Endosc 2008;67:791-8.

Moving To ensure continued service please notify us of a change of address at least 6 weeks before your move. Phone Subscription Services at 800-654-2452 (outside the U.S. call 314-4478871), fax your information to 314-447-8029, or e-mail [email protected].

954 GASTROINTESTINAL ENDOSCOPY Volume 73, No. 5 : 2011

www.giejournal.org