J Gastrointest Surg (2015) 19:1691–1698 DOI 10.1007/s11605-015-2868-2

ORIGINAL ARTICLE

Predictors of Successful Endoscopic Closure of Gastrointestinal Defects: Experience from a Single Tertiary Care Center Kamron Pourmand 1 & Brian Riff 2 & Michael L. Kochman 2 & Gregory G. Ginsberg 2 & Vinay Chandrasekhara 2 & Nuzhat A. Ahmad 2

Received: 12 April 2015 / Accepted: 28 May 2015 / Published online: 13 June 2015 # 2015 The Society for Surgery of the Alimentary Tract

Abstract Background/aims Endoscopic closure is technically feasible in the majority of gastrointestinal defects. The aim of this study is to evaluate the technical and clinical outcomes, and identify variables predicting successful outcomes in patients with attempted closure. Methods This is a retrospective study of patients undergoing endoscopic closure of gastrointestinal defects between December 2007 and May 2013 at a single tertiary care center. Technical success (TS) was defined as successful closure of the defect at the time of endoscopy. Clinical improvement (CI) was defined as improvement in symptoms. Clinical resolution (CR) was defined as documented radiographic closure of defect or clinical resolution of symptoms. Acute defects were diagnosed within 6 weeks, while chronic defects were those that persisted for >6 weeks, prior to index therapeutic endoscopy. Results Fifty patients underwent 77 endoscopies for leaks (n=23), fistulas (n=22), and perforations (n=5). TS occurred in 46/50 (92 %). Overall, 34/50 (68 %) patients had CR. CR was significantly higher for acute defects as compared to chronic defects (89.7 vs. 38.1 %, OR 14.1, CI 3.19–62.1, p6 weeks prior to the index therapeutic endoscopy were defined as chronic defects. Technical success was defined as successful endoscopic closure immediately at the conclusion of a procedure per the primary endoscopist’s narrative. Technical failure occurred whenever success was not evident at the time of endoscopy, or when partially successful closure was achieved. Initial clinical improvement (CI) was defined as either subjective improvement in symptoms (i.e., cough, drainage) or objective improvement (i.e., decreased leak on barium study) following an endoscopic closure attempt including but not requiring definitive resolution of symptoms. Clinical resolution (CR) was defined as definitive endoscopic closure at the time of follow-up, as determined by clinical and/or cross-sectional or barium studies (Fig. 2). Endoscopic procedures without discrete follow-up were excluded from analysis. The Institutional Review Board at the Hospital of the University of Pennsylvania (HUP) approved the study. Statistical Analysis Descriptive statistics consisted of means and standard deviations or medians with interquartile ranges for continuous variables and simple proportions for dichotomous variables. Univariate analysis was performed to evaluate for factors associated with technical success and CR. Variables analyzed

J Gastrointest Surg (2015) 19:1691–1698

a

b

Fig. 2 a Barium swallow study, pre-endoscopy: esophago-pleural fistula. b Barium swallow study, post-endoscopy: esophago-pleural fistula post-stent placement, demonstrating absence of leak

include age, gender, type of leak, and chronicity of leak. The groups were compared using chi-square/Fisher exact test analyses, simple linear regression or two-sample t tests. All statistical analyses were performed with STATA software (version 12.0; StataCorp, College Station, TX). Traditional levels of statistical significance were applied with p values less than 0.05 considered significant.

Results Fifty-seven patients who underwent a total of 77 procedures were identified as eligible for inclusion in our study. Four endoscopists (NAA, GGG, MLK, VC) performed all the procedures. Seven patients were excluded given the lack of meaningful documented follow-up within the electronic medical record system. Fifty patients (28 males; mean age 58 years, range 23–80 years) constituted our study cohort. The mean

1693

number of endoscopies per patient was 1.98. The baseline clinical characteristics of the patients are presented in Table 1. There were almost an equal number of patients with leaks (n= 23, 46 %) and fistulas (n=22, 44 %). At the time of the index endoscopic closure, the majority of leaks (14/23, 61 %) were acute, whereas the majority of fistulas were chronic (12/22, 54 %). For all closures, the most commonly utilized modality was endoscopic stent placement (n=30, 39 %) followed by OTSC (n=12, 15.6 %). The most common multi-modality approach in a single setting was with through-the-scope clips and fibrin glue (n=8, 10.4 %). Clinical or radiographic followup after the last endoscopy was available in 45 (90 %) patients, with a mean follow-up time of 119.2 days (range 1–854 days). The clinical course of our study population is outlined in Fig. 3. Overall, 34/50 (68 %) patients had CR after endoscopic closure. Technical success at conclusion of endoscopy was achieved in 46/50 (92 %). Of these patients, 29 (63 %) reported improvement in symptoms or had documented improvement in defect (CI), while 17 (37 %) were clinical failures. Of the 29 patients with CI, 20 (69 %) had CR. Nine patients developed recurrent symptoms and required a repeat attempt at endoscopic closure, with subsequent CR in all 9 patients. Of the 17 patients with initial clinical failure, 12 underwent a repeat endoscopy, with subsequent CR in 5 (42 %) patients. There were four patients who did not have technical success on their first endoscopy attempt, three of which were chronic leaks. The details of these patients are as follows: a patient with acute TE fistula post-transhiatal esophagectomy had a failed attempt at OTSC closure, ultimately requiring a bronchial stent (Merit Bronchial stent 12×30 mm) for closure. The second patient was status post Roux-en-Y gastric bypass with a fistula at the surgical gastroplasty. This patient had initially failed through-the-scope clip closure and later underwent surgical repair. The third patient had a postsurgical colo-anal anastomotic fistula. This patient failed fibrin glue placement and required a surgical anal fistulectomy for CR. The fourth patient with a gastro-gastric fistula experienced technical failure with through-the-scope clips, but subsequently had spontaneous CR of their defect. There were five patients with esophageal ruptures, who comprised the most acute cohort of defects in our study. Two of these patients had endoscopic intervention within 24 h with CR. The details of the other patients without Bimmediate^ intervention are as follows: patients had endoscopic therapy from 19–37 days post-esophageal perforation. Two of these patients achieved CR. Predictors of Technical Success Overall technical success (TS) at conclusion of endoscopy was achieved in 46/50 (92 %) patients. TS was achieved in 96.7 % (n=29/30) of patients where stent placement was

1694

J Gastrointest Surg (2015) 19:1691–1698

Table 1 Baseline characteristics of patients

Mean age, years (range)

58 (23–80)

Sex (male, %) Types of defects

28 (56 %) Acute, n (%)

Chronic, n (%)

No. of endoscopic procedures

23

14

9

37

13 7

10 3

3 4

18 16

3 22

1 10

2 12

3 33

Tracheoesophageal

11

10

1

18

Enterocutaneous Large bowel

7 3

0 0

7 3

10 4

Gastro-gastric Esophageal ruptures

1 5

0 5

1 0

1 7

Leaks Anastomotic Post-gastric leeve Post-RYGB Fistulas

RYGB Roux-en-Y gastric bypass

the sole modality for endoscopic closure and in 83.3 % patients where OTSC (n=10/12) was the sole closure modality. Through-the-scope clips alone (7/11) or in combination with stent (3/5) had the lowest TS rate (63.6 and 60 %, respectively). TS as classified by defect type are outlined in Table 2. The lowest rate of TS was seen in patients with RYGB and colon fistulas (66.6 % each). There was no statistically significant association between TS and type of defect, with technical success seen in

95.6 % of leaks, 86.3 % fistulas, and 100 % of esophageal perforations (p = 0.58). When evaluating patients with acute versus chronic defects, there was no statistically significant difference in TS rates between the two groups; there were 28/29 (96.6 %) patients with acute defects who experienced technical success, compared to 18/21 (85.7 %) patients with chronic defects (p=0.297). Technical success was also not significantly associated with age (p=0.17) or gender (p=0.80). Patients (n=57)

CR (n=9)

Excluded (n=7) Endoscopic closure (n=50)

Surgery (n=3)

Repeat Endoscopy (n=3)

Technical Failure (n=4)

Spontaneous Closure (n=1)

TS (n=46)

Repeat Endoscopy (n=9)

CI (n=29)

CR (n=20) Clinical Failure (n=17)

Surgery (n=3)

No further intervention (n=1)

Repeat Endoscopy (n=12)

No CR (n=7)

CR (n=5) TS = Technical Success CI = Clinical Improvement CR = Clinical Resolution

Fig. 3 Flow diagram of study cohort

Spontaneous Closure (n=1)

Surgery (n=2)

No CR (n=5)

J Gastrointest Surg (2015) 19:1691–1698 Table 2 type

1695

Technical success and clinical resolution classified by defect

Table 3

Clinical outcomes based on defect acuity

Defect type (total patients) Defect type (total patients)

TS (%)

CR (%)

Leaks (n=23)

22 (95.6)

15 (65.2)

Anastomotic (n=13)

13 (100)

10 (76.9)

Post-gastric sleeve (n=7) Post-RYGB (n=3)

7 (100) 2 (66.6)

3 (42.9) 2 (66.6)

19 (86.3)

15 (68.2)

Tracheoesophageal (n=11) Enterocutaneous (n=7)

10 (90.1) 7 (100)

9 (81.2) 5 (71.4)

Large bowel (n=3) Gastro-gastric (n=1)

2 (66.6) 0 (0)

1 (33.3) N/A

Esophageal ruptures (n=5)

5 (100)

4 (80)

Overall (n=50)

46 (92)

34 (68)

TS (%)

CR (%)

14 (100)

13 (92.3)

10 (100) 3 (100)

9 (90) 3 (100)

1 (100) 9 (90)

1 (100) 9 (90)

9 (90)

9 (90)

5 (100) 28 (96.6)

4 (80) 26 (89.7)

Acute defects

Fistulas (n=22)

RYGB Roux-en-Y gastric bypass, TS technical success, CR clinical resolution

Leaks (n=14) Anastomotic (n=10) Post-gastric sleeve (n=3) Post-RYGB (n=1) Fistulas (n=10) Tracheoesophageal (n=10) Esophageal ruptures (n=5) Overall (n=29) Chronic defects Leaks (n=9)

8 (88.9)

2 (22.2)

Anastomotic (n=3)

3 (100)

1 (33.3)

Post-gastric sleeve (n=4) Post-RYGB (n=2)

4 (100) 1 (50)

0 (0) 1 (50)

10 (83.3) 1 (100)

6 (50) 0 (0)

7 (100)

5 (71.4)

2 (66.6) 0 (0) 18 (85.7)

1 (33.3) N/A 8 (38)

Fistulas (n=12) Tracheoesophageal (n=1)

Predictors of Clinical Resolution Overall, 34/50 (68 %) patients had CR after endoscopic closure. There were 20/50 (40 %) patients who had CR after the first endoscopic closure and 14/50 (28 %) patients who achieved CR after more than one attempt at endoscopic closure. Though there was a trend towards an association between TS and CR, this did not reach statistical significance (p=0.09). The clinical outcomes as classified by defect type and defect acuity are detailed in Tables 2 and 3. Among the patients with acute defects, esophageal perforations had the lowest rate of CR (80 %). None of the chronic gastric sleeve leaks (n=4) or the chronic TE fistula (n=1) had CR. Overall, acute defects were statistically more likely to achieve CR as compared to chronic defects; 26/29 (89.7 %) patients with acute defects had CR, as compared to 8/21 (38 %) patients with chronic defects and CR (OR 14.1, CI 3.19–62.1, p6 weeks prior to closure. This effect can be explained by the fact that in the acute setting, given the absence of fibrotic scar tissue, the tissue planes are more amenable to endoscopic closure. Chronic defects, on the other hand, are more complex to manage, because abscesses or inflammation at the site with subsequent fibrosis has typically complicated these. Chronic defects also do not lend themselves as easily to closure with glue or stents because of epithelialization of the tract. Epithelial ablation or abrasion of chronic defects is critical for successful closure. Third, we also observed that nearly half of our patients underwent more than one attempt at endoscopic closure. The majority of these had CR, which emphasizes the point that a repeat attempt should be made in cases when indicated. We did note, however, that the response rate was higher in patients who had an initial improvement in symptoms or radiological appearance. It could be hypothesized that these defects had at least partially closed, resulting in easier targets for repeat intervention. The rate of complications in our study, most notably stent migration, was comparable to the 19 to 40 % rates reported in , – the literature.15 26 28 Stent migration is more likely with the use of fully covered stents, particularly given the lack of tethering of stent by the absence of luminal stenosis. Our study had several limitations. These included limitations inherent to retrospective studies. We were unable to ascertain variables such as leak size, which can affect success, from review of the records.29 In addition, the overall population of our study was very heterogeneous, and therefore, the cohort for each specific closure method, especially when tied to a specific type of defect, was small, which limited the power of our data. However, the comparability of patients with similar types of defects or endoscopic closure methods will always be problematic given the variability in the individual underlying health status of patients. Our study was also performed at a single tertiary care center and, hence, is subject to referral bias, as well as lack of generalizability to the wider community. However, because of the complexity of certain defects, we do recommend that these be treated at centers where there is expertise in their management. Finally, newer

1697

closure modalities such as the OTSC are more likely to be the device of choice for specific defects in the future as opposed to the timeframe included in this study. Despite these weaknesses, our study provides insight into variables that predict successful outcomes after endoscopic closure. In conclusion, our study supports the trend towards endoscopic management of gastrointestinal defects. The vast majority of defects can be safely closed, using a variety of devices, with reasonable outcomes. Early recognition and timely endoscopic management of a defect appears to be the critical factor with impact on outcomes.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

Golub R, Golub RW, Cantu R Jr, Stein HD. BA multivariate analysis of factors contributing to leakage of intestinal anastomoses.^ J Am Coll Surg. 1997 Apr;184(4):364-72. Schweigert M, Dubecz A, Beron M, Muschweck H, Stein HJ. BManagement of anastomotic leakage-induced tracheobronchial fistula following oesophagectomy: the role of endoscopic stent insertion.^ Eur J Cardiothorac Surg. 2012 May;41(5):e74-80. Freeman RK, Ascioti AJ. BEsophageal stent placement for the treatment of perforation, fistula, or anastomotic leak.^ Semin Thorac Cardiovasc Surg. 2011 Summer;23(2):154-8. Lippert E, Klebl FH, Schweller F et al. BFibrin glue in the endoscopic treatment of fistulae and anastomotic leakages of the gastrointestinal tract.^ Int J Colorectal Dis. 2011 Mar;26(3):303-11. Qadeer MA, Dumot JA, Vargo JJ et al. BEndoscopic clips for closing esophageal perforations: case report and pooled analysis.^ Gastrointest Endosc. 2007 Sep;66(3):605-11 Bhardwaj A, Cooney RN, Wehrman A et al. BEndoscopic repair of small symptomatic gastrogastric fistulas after gastric bypass surgery: a single center experience.^ Obes Surg. 2010 Aug;20(8): 1090-5. Parodi A, Repici A, Pedroni A et al. BEndoscopic management of GI perforations with a new over-the-scope clip device (with videos).^ Gastrointest Endosc. 2010 Oct;72(4):881-6. Weiland T, Fehlker M, Gottwald T, Schurr MO. BPerformance of the OTSC System in the endoscopic closure of iatrogenic gastrointestinal perforations: a systematic review.^ Surg Endosc. 2013 Jan 24. Baron TH, Song LM, Ross A et al. BUse of an over-the-scope clipping device: multicenter retrospective results of the first U.S. experience (with videos).^ Gastrointest Endosc. 2012 Jul;76(1): 202-8 Arezzo A, Verra M, Reddavid R et al. BEfficacy of the over-thescope clip (OTSC) for treatment of colorectal postsurgical leaks and fistulas.^ Surg Endosc. 2012 Nov;26(11):3330-3. Haito-Chavez Y, Law JK, Kratt T et al. International multicenter experience with an over-the-scope clipping device for endoscopic management of GI defects (with video). Gastrointest Endosc. 2014 Oct;80(4):610-22. Fischer A, Schrag HJ, Goos M et al. BNonoperative treatment of four esophageal perforations with hemostatic clips.^ Dis Esophagus. 2007;20(5):444-8. Sedarat A, Ginsberg GG, Ahmad N. An over-the-scope clipping device inadvertently sealing the pylorus: first description of a removal method. Gastrointest Endosc. 2014 May;79(5):711.

1698 14.

Dua KS, Vleggaar FP, Santharam R, Siersema PD. Removable selfexpanding plastic esophageal stent as a continuous, non-permanent dilator in treating refractory benign esophageal strictures: a prospective two-center study. Am J Gastroenterol. 2008 Dec;103(12):298894. 15. Seven G, Irani S, Ross AS et al. Partially versus fully covered selfexpanding metal stents for benign and malignant esophageal conditions: a single center experience. Surg Endosc. 2013 Jun;27(6): 2185-92. 16. Schweigert M, Solymosi N, Dubecz A et al. Endoscopic stent insertion for anastomotic leakage following oesophagectomy. Ann R Coll Surg Engl. 2013 Jan;95(1):43-7. 17. Shin JH, Song HY, Ko GY et al. Esophagorespiratory fistula: longterm results of palliative treatment with covered expandable metallic stents in 61 patients. Radiology. 2004 Jul;232(1):252-9 18. Dumonceau JM, Cremer M, Lalmand B, Devière J. Esophageal fistula sealing: choice of stent, practical management, and cost. Gastrointest Endosc. 1999 Jan;49(1):70-8. 19. May A, Ell C. Palliative treatment of malignant esophagorespiratory fistulas with Gianturco-Z stents. A prospective clinical trial and review of the literature on covered metal stents. Am J Gastroenterol. 1998 Apr;93(4):532-5. 20. Kozarek RA, Raltz S, Brugge WR et al. Prospective multicenter trial of esophageal Z-stent placement for malignant dysphagia and tracheoesophageal fistula. Gastrointest Endosc. 1996 Nov;44(5): 562-7. 21. Dasari BV, Neely D, Kennedy A et al. The role of esophageal stents in the management of esophageal anastomotic leaks and benign esophageal perforations. Ann Surg. 2014 May;259(5):852-60.

J Gastrointest Surg (2015) 19:1691–1698 22.

23.

24.

25.

26.

27.

28.

29.

Eubanks S, Edwards CA, Fearing NM et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg. 2008 May;206(5):935-8 Serra C, Baltasar A, Andreo L et al. Treatment of gastric leaks with coated self-expanding stents after sleeve gastrectomy. Obes Surg. 2007 Jul;17(7):866-72. Oshiro T, Kasama K, Umezawa A et al. Successful management of refractory staple line leakage at the esophagogastric junction after a sleeve gastrectomy using the HANAROSTENT. Obes Surg. 2010 Apr;20(4):530-4. Sakran N, Goitein D, Raziel A, et al. BGastric leaks after sleeve gastrectomy: a multicenter experience with 2,834 patients.^ Surg Endosc. 2013 27:240-245. van Boeckel PG, Dua KS, Weusten BL et al. Fully covered selfexpandable metal stents (SEMS), partially covered SEMS and selfexpandable plastic stents for the treatment of benign esophageal ruptures and anastomotic leaks. BMC Gastroenterol. 2012 Feb 29;12:19. Gelbmann CM, Ratiu NL, Rath HC et al. Use of self-expandable plastic stents for the treatment of esophageal perforations and symptomatic anastomotic leaks. Endoscopy. 2004 Aug;36(8):695-9. Sharma P, Kozarek R; Practice Parameters Committee of American College of Gastroenterology. Role of esophageal stents in benign and malignant diseases. Am J Gastroenterol. 2010 Feb;105(2):25873 Hajj I, Imperiale T, Rex D, et al. BTreatment of esophageal leaks, fistulae, and perforations with temporary stents: evaluation of efficacy, adverse events, and factors associated with successful outcomes.^ Gastrointest Endosc. 2014 Apr;79(4):589-98.