Omental plugging for large-sized duodenal peptic perforations: a prospective randomized study of 100 patients.

Background: Due to friable margins and the moribund state of the patient, managing giant duodenal perforations (>20 mm in diameter) is a challenging task.

Methods: A prospective randomized study of 100 patients with large-sized (> 20 mm) duodenal peptic perforation comparing omental plugging (study group) with omentopexy (control group) was carried out.

Results: Size of the perforation varied between 20 to 30 mm. No study group patients developed a postoperative perforation site leak, as compared with 6 patients in the control group. Gastric outlet obstruction was significantly less at 6 weeks and 5 years in the study group as compared with the control group, and mortality was significantly less in the study group.

Conclusion: It was concluded that omental plugging was a safe and reliable method of treatment for large-sized duodenal peptic perforations.

Key Words: giant duodenal perforation, omental plugging, randomized study

**********

Perforation is one of the most catastrophic complications of peptic ulcer. In spite of modern advances in surgical, anesthetic and ancillary facilities, it still assumes life-threatening dimensions. A variety of surgical techniques have been advocated for the management of peptic perforation. However, these techniques are not without their drawbacks, especially while managing giant perforations (ie, > 20 mm in diameter). Mortality rates of up to 18% have been reported while managing large-sized duodenal perforations with the standard techniques. Thus, there is a need to find, evaluate and apply better methods of managing these catastrophes.

We present a prospective study carried out at the Department of Surgery, SSG Hospital and Medical College, Baroda, of 100 patients with giant (ie> 20 mm) peptic perforations managed by omental plug technique as advocated by Karanjia et al. (1)

We have compared this technique with one of the standard operations for peptic perforation: Graham's technique of omentopexy. (2) The primary endpoint of the study was postoperative 30-day mortality. The secondary endpoints were postoperative suture line leak, postoperative hemorrhage, evidence of gastric outlet obstruction by endoscopy at 6 weeks, evidence of gastric outlet obstruction by barium meal study at 6 months and evidence of gastric outlet obstruction by endoscopy at 5 year follow-up visit.

Prior approval of the hospital ethics committee was acquired before initiating the study. Prior consent of all patients was taken before entering them into the study.

Materials and Methods

A randomized prospective study of 100 patients was carried out at the Department of Surgery, SSG Hospital, Baroda, from 1/1/93 to 4/3/98. The patient population included all peptic duodenal perforations with size of perforation > 20 mm in diameter as assessed intraoperatively. The patients were broadly divided into 2 groups:

1. Study group (SG): These patients underwent the omental plug technique as advocated by Karanjia et al (1)

2. Control group (CG): These patients underwent Graham's technique of omentopexy. (2)

All the patients with giant duodenal perforation were randomly allocated to either group using random number tables, ensuring randomization of extraneous parameters. Patients from both groups received the same standard pre- and postoperative procedures and medications, including antibiotics and postoperative H2 blockers.

Peritoneal exudate swab was taken for culture-sensitivity testing. After confirming the lesion, the peritoneal soiling was cleansed with normal saline wash, and the fluid was suctioned off. The operating surgeon was then informed which procedure was to be performed, according to the randomization.

The technique of omental plugging was performed on the study group as follows: The tip of the intragastric tube was guided through the perforation. The free edge of the omentum was taken and sutured to the tip of the Ryle tube using 2 to 0 plain catgut. The Ryle tube was gently withdrawn, pulling the plug of omentum into the stomach. Approximately a 5 to 6 cm length of omental plug sufficed to occlude the perforation. The omentum was then fixed to the perforation site with 5 to 6 interrupted sutures of 2 to 0 chromic catgut taken between the omentum and the healthy duodenum, approximately 3 to 4 mm away from the margins of the perforation. An abdominal drain was placed.

A Gastrografin study was carried out on the 7th postoperative day. In the event of a leak, the patient was managed either conservatively or underwent a second operation. The patients were discharged after prescribing a course of famotidine 40 mg/d for 2 months. They were called for follow-up after 7 days, 15 days, 1 month, 2 months, 6 months and yearly thereafter. Upper GI endoscopy was done at 6 weeks to determine the condition of the ulcer site and whether gastric outlet obstruction was present. Rapid urease test was done at the same time to detect the presence of H. pylori. All patients positive for the organism were treated for four weeks on a triple regimen consisting of amoxicillin, metronidazole and omeprazole. Barium meal was done at 6 months. Upper GI endoscopy and barium series were repeated at the end of five years. Both the endoscopist and the radiologist were blinded to the procedure performed. Informed consent was taken for all the procedures.

The following parameters of the 2 groups were compared: age distribution, sex distribution, associated diseases, hemoglobin level, culture and sensitivity of peritoneal exudate, culture and sensitivity of blood, average hospital stay, morbidity/mortality, upper GI endoscopy at 6 weeks, barium meal study at 6 months and upper GI endoscopy and barium meal study at 5 years.

Results and Analysis

The patient demographics are detailed in Tables 1 and Table 2. The age of the patients varied from 16 years to 66 years with the average age in the study group being 39.2 years and 36.8 years in the control group. Males predominated, comprising 88% of the study group and 90% of the control. Both groups had statistically similar distributions concerning the variable of age and sex. The prevalence of associated diseases like diabetes mellitus, hypertension, ischemic heart disease, tuberculosis and acute renal failure was comparable in both the groups as described in Table 3.

The mean hemoglobin in the study group (SG) was 10 g/dL in males and 9.4 g/dL in females and 10.8 g/dL in males and 8.5 g/dL in females in the control group (CG). Forty percent of the SG and 44% of the CG had a positive culture of peritoneal fluid, and 16% each of the SG and CG had a positive blood culture. The predominant organisms were E. coli, Proteus and Klebsiella. All of the patients had a perforated ulcer located on the anterior aspect of the first part of duodenum, with a size varying between 20 to 30 mm. (Table 4). Both groups were comparable for these parameters.

Six patients in the CG developed a leak as evidenced by drainage > 100 mL/24 hours, which was confirmed by a Gastrografin study. All six of these patients required a second operation. None of the SG patients developed this complication. Reclosure was done utilizing the serosal patch technique, but all 6 patients died in the early postoperative phase due to septicemia.

Postoperative hemorrhage in the form of hemorrhagic Ryle tube aspiration occurred in 8% of the SG and 4% of the CG. All the patients were managed conservatively with blood transfusions and vital sign monitoring and all survived. The postoperative mortality in the SG was 8%, compared with 16% in the CG. All 4 of the SG patients who died had a history of abdominal catastrophe at least 48 hours old, and the cause of death was septicemia. All 8 of the mortalities in the CG were due to septicemia, out which 6 patients had a leak that required re-operation. The average hospital stay was 8 days for the SG and 8.8 days for the CG.

Upper GI endoscopy carried out at 6 weeks revealed partial outlet obstruction in 12% of the CG patients, but in none of the SG patients. The patients complained of vomiting partially digested food and on endoscopy, there was narrowing at the operation site. The patients were managed conservatively with small, frequent semisolid meals. Barium meal study at 6 months revealed outlet obstruction in 10% of the SG and 24% of the CG. All the patients underwent endoscopic balloon dilation, with symptomatic relief following 2 to 3 treatments.

At the end of 5 years, 64 of the patients could be followed up, of which 36 belonged to the SG and 28 belonged to the CG. None of the SG patients had any evidence of outlet obstruction; while 3 patients (10.7%) of the CG had features of obstruction. (Table 5) All of them underwent surgical correction with side-to-side retrocolic isoperistaltic gastrojejunostomy and bilateral truncal vagotomy.

Hence, the study group had a statistically significantly lower leakage rate and performed significantly better as far as gastric outlet obstruction was concerned, compared with the control group. There was no significant difference for other outcome parameters, such as postoperative hemorrhage, gastric outlet obstruction at 6 months, or mortality.

Discussion

There is a paucity of data in the literature regarding giant duodenal ulcers with some case reports and very few series. There has not been a single randomized control study for the management of this severe variant of duodenal ulcer disease. One of the reasons for this is that giant duodenal ulcers are an uncommon entity. In a series of 1,434 patients with peptic ulcers, giant duodenal ulcers were found in 2.4%. (3)

Eighty-four percent of our patients belonged to the 3rd, 4th, and the 5th decades, which is consistent with the available data. (4) The male: female ratio of 8.1:1 in our study corresponds with the available data. (4) In his series of 32 patients, Nussbaum (5) found that 75% of the patients were men between 30 and 81 years of age (mean age 59 yr). Gastrointestinal hemorrhage was a presenting symptom in 75% of the patients and free perforation in 9%. Unusual presentation, such as penetration into the liver, has also been reported. (6) The presence of other systemic diseases in our series was 4 to 10% in both groups, with the most common being hypertension and diabetes mellitus. The most common systemic diseases reported in Western countries are hypertension, diabetes and ischemic heart disease. (7-10) Nearly half of our patients had a past history suggestive of acid peptic disease more than 3 months duration. This figure is comparable to that available in literature. (8) The average hemoglobin content in our series was 10 g/dL which was low as compared with Western counterparts. (9,10)

The anemia-hypoproteinemia complex may have played a role to some extent in the incidence of leakage from the operative site on the duodenum and the high mortality rate due to septicemia, owing to the body's poor defense mechanism. Peritoneal fluid culture was positive in 42% and blood culture was positive in 16% of our patients, which is high compared with the 33.5% positive peritoneal fluid culture reported by Boey et al. (11) One of the reasons may be a late presentation of the patients to the hospital. This may in turn contribute to the high mortality due to septicemia. Although it has been suggested that uncomplicated giant duodenal ulcers can be managed with proton pump inhibitors, (12) surgery is the mainstay of treatment. These patients do very poorly with medical therapy, and the mortality rate is increased if emergency surgery is required. Medical treatment alone is associated with a high morbidity of around 92%. A definitive acid reduction operation is the procedure of choice. (5)

Several elaborate surgeries have been devised to manage complicated giant duodenal ulcers. Wu et al (13) performed a partial gastrectomy and duodenostomy, and a duodenostomy tube was placed through the duodenal stump, enveloped around an omental patch, for bleeding giant duodenal ulcers. Cranford et al (14) suggested a two-stage surgery in the form of truncal vagotomy and antrectomy, along with tubal gastrostomy, duodenostomy, and jejunostomy, followed by restoration of gastrointestinal continuity after 3 to 4 weeks.

In contrast to all these elaborate measures, the omental plug is a simple procedure which does not require significant expertise and can even be performed in a very short time by a trainee general surgeon in a seriously ill patient in an emergency situation. (15) In our series, whereas 6 patients in the CG had a leak from the duodenal closure site, none of the patients in the SG had this complication, suggesting that omental plugging is a very effective technique for treatment of giant duodenal perforation. Although the incidence of postoperative hemorrhage was higher in the SG (8% as compared with 4% of the CG), it could be managed conservatively and none of the patients of either group died to it. There was no operative mortality in either group. The early postoperative mortality in the SG was 8%, whereas in the CG, it was 16%. All of the SG patients who died had presented at least 48 hours after the initial abdominal catastrophe (onset of severe sudden abdominal pain and vomiting) and had died due to septicemia. However, 6 of the 8 patients in the CG who died were early presenters, but developed a leak at the duodenal operative site, and all died to septicemia. Literature reports of mortality rates varies from 2 to 18% for simple closure and 1.1 to 8% for definitive operations. (9,16-18)

The average hospital stay of 8 to 9 days in our series corresponds to that found in the literature. (16)

During follow-up, 12% of the CG patients revealed gastric outlet obstruction at 6 weeks on upper gastroesophageal endoscopy, while none of the SG patients had these complaints. Barium meal at 6 months revealed 10% of the SG, and 24% of the CG had gastric outlet obstruction. The difference in outlet obstruction is statistically highly significant at 6 weeks but is not significant at 6 months, implying that omental plugging may have better results as far as outlet obstruction is concerned than the standard technique in the short term, but both the techniques are comparable for medium term outcome of gastric outlet obstruction. However, at the five years follow-up, all patients that were available (n = 36) were free from gastric outlet obstruction, while 3 of the 28 patients belonging to the CG and available for follow-up had features of obstruction. Read (19) reported a 36% incidence of gastric outlet obstruction following omentopexy. Experimental evidence already exists testifying to the reliability of the omental plug in safely occluding large duodenal defects and producing healing through a process of inflammation, granulation, vascularization and fibrosis, eventually providing a normal duodenal mucosal cover to the perforation site. (20) In fact, a modification of the technique to adapt it to the minimally invasive approach of laparoscopy has already been reported. (21)

Thus, in the long term, omental plugging was superior to omentopexy in regards to gastric outlet obstruction.

Conclusion

The technique of omental plugging for large duodenal perforations is safe and fast and can be carried out in poor-risk patients. In our study, this technique had no operative mortality. Omental plugging can be safely performed with giant perforated duodenal ulcers with indurated, fibrotic, friable margins. The technique of omentopexy, which is the most commonly practiced method of closure for duodenal perforations, has a fairly high incidence of re-leak from the operated duodenal site as compared with omental plugging. There was decreased incidence of gastric outlet obstruction as compared with omentopexy. Performing a more definitive procedure like partial gastrectomy is more time consuming and is associated with a higher morbidity and mortality.

Thus, for giant duodenal peptic perforations, omental plugging is a safe and reliable method of treatment, especially in high-risk patients.

References

1. Karanjia ND, Shanahan DJ, Knight MJ. Omental patching of a large perforated duodenal ulcer: a new method. Br J Surg 1993;80:65.

2. Graham RR. Treatment of perforated duodenal ulcers. Surg Gynecol Obstet 1937;64:235.

3. Csendes A, Becker P, Valenzuela J, et al. Clinical characteristics of patients with multiple or giant peptic ulcers. Rev Med Chil 1991;119:38-44.

4. McKay AJ, McArdle CS. Cimetidine and perforated peptic ulcer. Br J Surg 1982;69:319-320.

5. Nussbaum MS, Schusterman MA. Management of giant duodenal ulcer. Am J Surg 1985;149:357-361.

6. Novacek G, Geppert A, Kramer L, et al. Liver penetration by a duodenal ulcer in a young woman. J Clin Gastroenterol 2001;33:56-60.

7. Sawyers JL, Herrington JL, Mulherrin Jr, JL et al. Acute perforated duodenal ulcer: an evaluation of surgical management. Arch Surg 1975;110:527-530.

8. Bahnini J, Kleiber G, Karidakis P, et al. Current status of the treatment of perforated duodenal ulcer. J Chir (Paris) 1985;122:121-128.

9. Hamby LS, Zweng TN, Strodel WE. Perforated gastric and duodenal ulcer: an analysis of prognostic factors. Am Surg 1993;59:319-323.

10. Boey J, Choi SK, Poon A, et al Risk stratification in perforated duodenal ulcers: a prospective validation of predictive factors. Ann Surg 1987;205:22-26.

11. Boey J, Wong J, Ong GB. Bacteria and septic complications in patients with perforated duodenal ulcers. Am J Surg 1982;143:635-639.

12. Fischer DR, Nussbaum MS, Pritts TA, et al. Use of omeprazole in the management of giant duodenal ulcer: results of a prospective study. Surgery 1999;126:643-649.

13. Wu X, Zen D, Xu S, et al. A modified surgical technique for the emergent treatment of giant ulcers concomitant with hemorrhage in the posterior wall of the duodenal bulb. Am J Surg 2002;184:41-44.

14. Cranford Jr, CA Olson R, Bradley EL. Gastric disconnection in the management of perforated giant duodenal ulcer. Am J Surg 1988;155:439-442.

15. Sharma D, Saxena A, Rahman H, et al. 'Free omental plug': a nostalgic look at an old and dependable technique for giant peptic perforations. Dig Surg 2000;17:216-218.

16. Braun L. Surgical therapy of ulcer disease: early and late results of elective and emergency interventions. Chirurg 1991;62:681-685.

17. Christiansen J, Andersen OB, Bonnesen T, et al Perforated duodenal ulcer managed by simple closure versus closure and proximal gastric vagotomy. Br J Surg 1987;74:286-287.

18. Jordan PH Jr. Indications for parietal cell vagotomy without drainage in gastrointestinal surgery. Ann Surg 1989;210:29-41.

19. Read RC, Thompson BW. Gastric outlet obstruction after omentopexy for perforated 'acute' and 'chronic' duodenal ulceration. Am J Surg 1975;130:682-687.

20. Raj BR, Subbu K, Manoharan G. Omental plug closure of large duodenal defects: an experimental study. Trop Gastroenterol 1997;18:180-182.

21. Pescatore P, Halkic N, Calmes JM, et al. Combined laparoscopic-endoscopic method using an omental plug for therapy of gastroduodenal ulcer perforation. Gastrointest Endosc 1998;48:411-414.

There is no one alive who is Youer than You.
--Dr. Seuss

Kalpesh Jani, MD, A.K. Saxena, MD, and Rasik Vaghasia, MD

From the Department of Surgery, Medical College & SSG Hospital, Baroda, India.

Reprint requests to Dr. Kalpesh Jani, Gem Hospital, 45A, Pankaja Mill Road, Ramanathapuram, Coimbatore-641045. Email: [e-mail]

Accepted October 7, 2005.

RELATED ARTICLE: Key Points

* Giant duodenal ulcers (>20 mm in diameter) are potentially life threatening, with medical therapy alone associated with a morbidity of around 92%.

* Perforation of these ulcers is a surgical catastrophe. Standard technique of omentopexy is associated with high morbidity and mortality.

* Omental plugging is a simple procedure which does not require significant expertise and can be performed emergently in seriously ill patients by a trainee general surgeon.

* It has lower short-term and long-term morbidity as compared with the standard technique of omentopexy for perforation closure.

Table 1. Age distribution

Age (years)  SG  CG  Total

< 10          0   0    0
11-20         0   1    1
21-30         4   9   13
31-40        24  24   48
41-50        20  13   33
51-60         1   2    3
61-70         1   1    2
Total        50  50  100

The standard error of difference between the two means (SED) is 1.67 and
the critical ratio (CR) is 1.44, which is not significant for P = 0.05.
SG, study group; CG, control group.

Table 2. Sex distribution

Sex      SG  CG  Total

Males    44  45   89
Females   6   5   11
Total    50  50  100

2 x 2 contingency [chi] square value = 0.10, which is not significant at
P = 0.05.
SG, study group; CG, control group.

Table 3. Presence of comorbidity

Associated diseases     SG  CG  Total

Diabetes mellitus        7   3  10
Hypertension             5   4   9
Tuberculosis             3   4   7
Ischemic heart disease   5   3   8
Acute renal failure      0   4   4
Total                   20  18  38

The 2 X 2 contingency [chi] square value is 0.17, which is not
significant at P = 0.05.
SG, study group; CG, control group.

Table 4. Hemoglobin, peritoneal fluid culture and blood culture

Investigations                         SG    CG

Mean Hb content (gm/dL) (a)
  Males                                10    10.8
  Females                               9.4   8.5
Positive peritoneal fluid culture (b)  20    22
Positive blood culture (c)              8     8

(a) The CR for the mean hemoglobin content was 1.08 for males and 1.00
for females, both of which were not significant for P = 0.05.
(b) The [chi] square value for positive peritoneal fluid culture was
0.16, which was not significant for P = 0.05.
(c) The [chi] square value for positive blood culture was 0, both of
which were not significant for P = 0.05.
SG, study group; CG, control group.

Table 5. End points of the study

Parameter                                      SG           CG

Postoperative leakage from operated site (a)   0             6
Gastric outlet obstruction by upper GI         0             6
  endoscopy at 6 weeks (b)
Gastric outlet obstruction by barium meal at   5            12
  6 months (c)
Endoscopic evidence of gastric outlet          0 (of 36      3 (of 28
  obstruction at 5 years                         patients)     patients)
Postoperative hemorrhage (d)                   4             2
Postoperative mortality (e)                    4             8
Average hospital stay in days                  8             8.8

(a) The [chi] square value for leakage from the operated site was 6.38,
which was highly significant even at P = 0.01.
(b) The [chi] square value for gastric outlet obstruction as detected by
upper GI endoscopy at 6 weeks was 6.38, which was highly significant at
P = 0.01.
(c) The [chi] square value for gastric outlet obstruction as detected by
barium meal at six months was 3.47, which was not significant for P =
0.05.
(d) The [chi] square value for postoperative hemorrhage was 0.71, which
was not significant for P = 0.05.
(e) The [chi] square value for mortality was 1.52, which was not
significant for P = 0.05.
SG, study group; CG, control group.