Effect of an Omental Wrap on the Healing and Vascularity of Compromised Intestinal Anastomoses Warwick Adams, F.R.A.C.S., Graeme Ctercteko, F.R.A.C.S., M.D., Michael Bilous, F.R.C.P.A. From the Department of Surgery, Westmead Hospital, Westmead, NSW, Australia Adult Wistar rats were used to investigate the ability of an omental wrap to limit leakage from compromised intestinal anastomoses. Under ketamine anesthesia, a section of small bowel was divided and then reanastomosed using a "control" anastomosis, a "deficient" anastomosis, or an "ischemic" anastomosis, plus or minus the addition of a wrap of omentum. Initially 10 rats were randomly assigned to each group. Nineteen of the 20 rats with unwrapped compromised anastomoses died within six weeks, compared with five deaths in the rats protected by an omental wrap (Fisher's exact test; P < 0.01). The experiment was then repeated with a sample of rats from each anastomotic group being sacrificed for histologic examination on days 2 to 7, 10, 14, and 42. At the time of sacrifice a dye was injected into the omental vasculature to determine its contribution to the healing anastomosis. An anastomosis could be demonstrated between omental and bowel wall vessels by the third postoperative day. At one week the infarcted bowel edges were being resorbed and the omentum formed a fibrotic cylinder aligning the separated ends of bowel wall. At six weeks the scar became more contracted and the bowel mucosa had started to grow onto its luminal surface. It is concluded from this study that the omental wrap is protective to a compromised anastomosis by providing a biologically viable plug to prevent early leakage and a source of granulation tissue and neovasculature for later wound repair. [Key words: Intestinal anastomoses; Omental wrap; Neovasculature; Intraperitoneal adhesions] Adams W, Ctercteko G, Bilous M. Effect of an omental wrap on the healing and vascularity of compromised intestinal anastomoses. Dis Colon Rectum 1992 ;35:731 738. he p e d i c l e d omental graft has wide uses in surgery. Since Bennett first described its use in 1896 to plug a perforated gastric ulcer, 1 it has b e e n described for clinical use in the repair of vesicovaginal fistulas, 2 liver injuries, 3' 4 and b o d y wall defects, 5 urinary tract reconstruction, 6 and the protection of vascular grafts. 7 Some authors have described its clinical use for protection of gastrointestinal anastomoses, particularly after rectal s' 9 and esophageal anastomoses. 1~ O n our o w n unit it is customary to wrap low rectal
T
No reprints are available.
anastomoses in o m e n t u m to r e d u c e the risk of leakage. However, the experimental evidence to support this is contradictory, n 13 In this study we have used an animal m o d e l to investigate the effects of an omental wrap w h e n placed around a potentially leaking anastomosis.
MATERIALS AND METHODS Adult 300-g Wistar rats were used in the study. They were h o u s e d and bred in a vivarium in the grounds of the Westmead Hospital, u n d e r the care of veterinary staff e m p l o y e d by the University of Sydney. Approval to c o n d u c t this study was granted by the C u m b e r l a n d Area Health Service Research Committee. Three anastomotic techniques were investigated: a control anastomosis, w h i c h p r o d u c e d a reliable join, a mechanically deficient anastomosis, and an ischemic anastomosis. These latter two t e c h n i q u e s p r o d u c e d c o m p r o m i s e d anastomoses w h i c h simulated, in a s o m e w h a t exaggerated fashion, the anastomotic p r o b l e m s of mechanical failure and ischemia. The e x p e r i m e n t was then p e r f o r m e d in two parts. In the first part, the effect of an omental wrap on anastomotic c o m p e t e n c e was determined. The seco n d part of the e x p e r i m e n t investigated the histologic events occurring in the incorporation of o m e n t u m into a c o m p r o m i s e d anastomosis and the timing of the induction of a neovasculature.
Anastomotic Technique operations were p e r f o r m e d through
The a midline incision u n d e r intraperitoneal ketamine anesthesia, using a sterile technique. The animals were given access to only water and dilute lactulose solution for 24 hours preoperatively to clear the bowel and were elevated on a grid off the floor of their cage to prevent coprophagy. 731
ADAMS E T A L
732
The male rats have a large retractile testis with an associated pad of fat which is in fact larger than the omentum. This was found in early experiments to wrap itself around the anastomoses and so confuse the results. For this reason, this fat pad was routinely ligated and excised in all rats in the study. The mid small bowel was divided at the point of insertion of a mesenteric vessel and then reanastomosed. A continuous Connell suture with 7/0 silk was used for the control anastomosis. Two techniques were used to produce compromised anastomoses. The first was mechanically deficient in that only three interrupted sutures were used to oppose the bowel ends. The second was performed with four interrupted sutures and then made ischemic by ligating the mesenteric vessel and the marginal vessels 2 cm on either side of the anastomosis (Fig. 1). In the rat the omentum hangs free from the greater curve of the stomach and is not attached to the colon. To perform the omental wrap the anastomosed bowel was brought up to lie adjacent to the stomach. The omentum was then wrapped around the anastomosis to completely encircle it. It was maintained in position by sutures placed between the base of the omentum and the bowel
C.
Dis Colon Rectum, August 1992
mesentery behind the anastomosis and from the tip of the omentum to the bowel mesentery in front of the anastomosis (Fig. 2).
Determination of Anastomotic
Competence For the first part of the experiment, 50 rats underwent laparotomy and bowel anastomosis. Ten animals were randomly assigned to each of five experimental groups: a control group, a deficient anastomotic group, an ischemic anastomotic group, a deficient group with an omental wrap, and an ischemic group with an omental wrap. All rats that died underwent postmortem examination on the day of death. Survivors were sacrificed at six weeks. At postmortem examination the anastomoses were assessed for the presence of intraperitoneal leakage, perianastomotic abscess, and the formation of adhesions. The propensity to develop adhesions was assessed by measurement of the proportion of the circumference of the anastomosis adherent to surrounding organs.
a
be d.
Figure 1. Diagram of small bowel anastomoses, a. Bowel divided opposite mesenteric vessel, b. Control anastomosis; continuous suture with 7/0 silk. c. "Deficient anastomosis"; three interrupted sutures, d. "lschemic anastomosis"; four anastomotic sutures. The mesenteric vessels have been ligated 2 cm distant from either side of the anastomosis.
Figure 2. Diagram of omental wrap. a. Omentum sutured to bowel mesentery posteriorly, b. Circumferential wrap and omentum sutured anteriorly.
Vol. 35, No. 8
EFFECT OF AN ANASTOMOTIC OMENTAL WRAP
The anastomotic segment was then excised. It was fixed in formalin and prepared in a paraffin block. After staining with hematoxylin and eosin, it was submitted to histologic examination. Histologic Events For the second part of the experiment a sample of rats from each anastomotic group was sacrificed on selected postoperative days for submission to histologic study. The numbers sacrificed from each group, and the days of sacrifice, are shown in Table 1. Because the unwrapped anastomoses had a much higher mortality rate in the early postoperative period, it was found necessary to perform more of these anastomoses to produce sufficient numbers for later examination. To demonstrate the vascular contribution from the omentum to the healing anastomosis, a dye was injected into the omental vasculature at the time of the rat's sacrifice. This technique involved reanesthetizing the rat and exploring the abdomen. The anastomosed segment of bowel was isolated from the rest of the gut and its mesentery and was left attached only to the omental pedicle. The chest was then opened and the descending aorta cannulated. India ink was then injected until the omental Table 1. Timing of Postmortem Examination After Intestinal Anastomosis Number of Rats with each Anastomotic Technique
Day of Examination Ischemic Deficient Ischemic Deficient with with (no (no Wrap Wrap Wrap) Wrap) 1 2 3 4 5 7 10 14 18 21 42 Total
1" 2 2 2 2 2 2 1" 1" 1" 2 18
0 2 2 2 2 2 2 1 1" 0 2 16
12" 141" 2 1 0 0 0 1 0 0 0
9* 131" 2 1 1 0 1 0 0 0 0
30
27
All rats, except those marked with an asterisk, were sacrificed and underwent a dye study to outline the omental vasculature. * Spontaneous deaths, no dye injected. 1"Two sacrificed and dye study performed; remainder were spontaneous deaths.
733
vessels were seen to fill. The anastomosed segment of gut was then excised and examined. An analogous study was also performed on those rats sacrificed without an omental wrap in which spontaneous adhesions surrounded the anastomosis, the anastomosed segment of gut was isolated on a pedicle of adhesions, instead of omentum, and a similar injection technique was used. Under the microscope the dye could clearly be seen filling the omental vessels, including the capillaries. As capillaries budded into new vessels in the granulation tissue, these also filled with dye and their progress could be mapped. Statistical M e t h o d s Differences between anastomotic groups for both survival and the propensity to induce adhesions were compared using Fisher's exact test. The level for statistical significance was set at P < 0.05. RESULTS Survival In the first experiment there was a statistically significant survival advantage found with the addition of an omental wrap, in both deficient and ischemic anastomoses (Table 2). Fifteen of the 20 rats with either ischemic or deficient unwrapped anastomoses died in the first 48 hours. Inspection of the peritoneal cavity in each case revealed feculent peritonitis with obvious leaks from anastomotic gaps. Microscopic examination demonstrated infarcted bowel edges and an absence of bowel apposition. Three rats died after the first week as a result of bowel obstruction from anastomotic stricture. In each case a perianastomotic abscess was present, enclosed within surrounding adhesions. The addition of the omental wrap reduced the incidence of early fatalities but did not affect the incidence of delayed anastomotic stricture. There were three strictures in the deficient group with a wrap and two in the ischemic group with a wrap. In the deficient group, one stricture led to a fatal obstruction and two produced an incomplete obstruction in survivors at six weeks. In the rats with ischemic anastomoses and a wrap, two deaths followed early anastomotic leaks, one of which as a result of the omentum having been twisted during its preparation and rendered nonviable, and two deaths occurred later from obstruction.
ADAMS E T AL
734
Dis Colon Rectum, August 1992
Table 2. Number of Rats Surviving to Six Weeks Anastomotic Technique Used
Rats Dying Before Six Weeks
Rats Alive at Six Weeks
Total
Deficient Ischemic Deficient with omental wrap Ischemic with omental wrap Control
9* 10t 1" 4t 1
1* 0t 9* 6t 9
10 10 10 10 10
* P < 0.0005. t P < 0.005.
Table 3. Incidence of Anastomotic Adhesions (Measured as a Percentage of the Anastomotic Circumference Involved in Adhesions) Number of Rats with Anastomotic Adhesions
Anastomotic Technique Used
Deficient Ischemic Deficient with omental wrap Ischemic with omental wrap Control
No Adhesions
Adhesions Less Than 50% of Anastomosis
Adhesions More Than 50% of Anastomosis
0 0 7 6 0
0 0 2 2 0
10 10 1 2 9*
* One unable to be assessed.
There was one death in the control group. The cause of death was unclear since the anastomosis was identified within the remains and found to be intact. The other nine rats survived to six weeks with intact and patent joints. On microscopic examination these had healed by primary intention with inverted bowel edges and mucosal apposition.
Adhesions All anastomoses not wrapped in omentum developed adhesions to surrounding intraperitoneal structures. In all rats surviving beyond 48 hours, these were completely circumferential. In contrast to this, 13 of the 20 anastomoses wrapped with omentum formed no adhesions at all (Table 3). This difference between wrapped and unwrapped groups was statistically significant ( P < 0.001). In the seven rats with an omental wrap in which some perianastomotic adhesions were present, an inflammatory reaction extended through the full thickness of the overlying omentum. In the remaining 13 cases without adhesions, the inflammatory reaction was confined to the luminal surface of the omentum.
Patterns Observed in the Healing of Anastomoses with an Omental Wrap In the second part of the experiment, there were five deaths among the 34 wrapped anastomoses, one from an early leak and four from bowel obstruction. The ischemic preparation was found to consistently produce a 15- to 20-mm length of infarcted bowel at the anastomotic site. The deficient technique produced anastomotic gaps 3 to 5 mm in length. In the first 48 hours, the omentum became adherent to the ischemic bowel wall and bridged any anastomotic defects (Fig. 3). By day 3, granulation tissue was seen to develop on the luminal surface of the omentum. At this stage dye was seen to pass from omental to bowel wall vessels in all four cases studied (Figs. 4 and 5), demonstrating the induction of a vascular anastomosis between the omental blood supply and the anastomosed bowel. Over the next six weeks, the infarcted bowel edges were resorbed and collagen was laid down in the cylinder of omentum joining the separated ends of bowel. In those cases in which bowel obstruction
Vol. 35, No. 8
EFFECT OF AN ANASTOMOTIC OMENTAL WRAP
Figure 3. Ischemic anastomosis wrapped in omentum, at 48 hours. One side of the anastomosis is shown. The omentum can be seen spanning the gap between the infarcted bowel edges (marked by curved arrows). Acute inflammatory changes are present on the luminal surface of the omentum (shown by straight arrow). Hematoxylin and eosin, x29.
735
Figure 5. Higher power of Figure 4 to show dye within the bowel wall capillaries. Hematoxylin and eosin, x350.
Figure 6. Deficient anastomosis wrapped in omentum, at 14 days. There is maturation of the granulation tissue and opening up of vascular channels. Hematoxylin and eosin,
x26. Figure 4. Deficient anastomosis wrapped in omentum, at 72 hours. Granulation tissue has formed on the luminal surface of the omentum (orientated toward the top of the picture). Black dye can be seen within the bowel wall vessels (arrows), better seen on higher magnification in Figure 5. Hematoxylin and eosin, x40. developed, an exuberant granulation tissue response had arisen from the omentum and occluded the lumen. Vascular anastomoses persisted between the omentum and bowel wall (Fig. 6). At six weeks some mucosal proliferation became evident at the scar edges. Patterns Observed in the Healing of Anastomoses Without a n O m e n t a l W r a p Fifty-seven rats were examined after the construction of compromised anastomoses without the
addition of an omental wrap. Forty-four died in the first 48 hours. In each of these cases, obvious leaks were seen at postmortem examination, through anastomotic gaps not covered by adhesions. Thirteen rats with unwrapped anastomoses were sacrificed and examined between days 2 and 14. In each case fully circumferential adhesions were present at the anastomosis. The adhesions were formed predominantly from loops of small intestine. Often liver, pancreas, abdominal wall, and sometimes a small tongue of omentum also contributed to the adhesion complex. The pattern of healing in those rats surviving beyond 48 hours was similar to that seen in rats with an omental wrap. The adhesions acted as a plug to the anastomotic defects to prevent early leakage. The luminal surface of the adhesions then
ADAMS E T A L
736
developed an inflammatory reaction, followed by the production of the major portion of the granulation tissue in the wound. In the two rats examined on days 10 and 14, this granulation tissue had matured into a fibrous collagen plug. A vascular anastomosis between the adhesions and the bowel wall was demonstrated in three of the four rats examined on day 3 and in all rats examined subsequently. DISCUSSION Using a lethal anastomotic technique, the omental pedicle wrap has provided a dramatic improvement in experimental animal survival by maintaining intestinal continuity and integrity over a distance between two separated ends of bowel. The omental wrap forms an effective bridge over anastomotic defects in the first 48 hours and then provides the bulk of the granulation tissue in their subsequent healing. An anastomosis develops between the omental and bowel wall vessels as early as the third postoperative day to aid in anastomotic healing, but the input of omental blood arrives too late to avoid bowel edge infarction from the original ischemic injury. Random adhesions around a compromised anastomosis can behave in a similar fashion to an omental wrap, including the development of a neovasculature. However, these adhesions do not reliably form a seal in the first 48 hours. The survival advantage found in the present study is consistent with the experience of McLachlin and Denton ~ and Katsikas e t aL Ig who both tested the omental wrap on ischemic intestinal anastomoses in dogs, but contrasts with the experience of Carter and colleagues, 13 who tested both pedicled and free omental wraps on a much less lethal anastomotic model in the rabbit. The problem identified with wrapped anastomoses in these studies, as in our own, was the development of delayed anastomotic strictures and bowel obstruction. However, we have found that these strictures have resulted from early anastomotic failure With bowel separation and infarction. The stricture results from the profuse inflammatory response in the surrounding wrap, followed by an exuberant growth of granulation tissue. In these cases the absence of a wrap would have inevitably led to an early leak and peritonitis. The fact that so many animals with a wrapped compromised anas-
Dis Colon Rectum, August 1992
tomosis actually had an untroubled convalescence with an intact anastomosis at six weeks demonstmtes the wrap's protective ability. Another feature seen in association with a wrap was the reduced incidence of perianastomotic adhesions, not only in comparison with unwrapped compromised anastomoses, but in comparison with controls. This was found to be associated with the containment of the perianastomotic inflammation to the luminal surface of the omentum. The biologic viability of the wrap appears to be critical for its function. The one case in this series in which the wrap was accidentally devascularized led to its early failure. Devascularized omentum and other tissues have been previously studied for their potential to protect anastomoses. < 11, 13. 14 It has been found that the devascularized omentum is not of benefit and actually leads to a worse anastomotic outcome than no wrapping at all. This phenomenon was well demonstrated by Ravitch e t al. 15 when they induced a high leak rate from anastomoses isolated from surrounding adhesions by a wrap of polyethylene or glove rubber. Ellis le found that a free omental flap actually drew a blood supply away from underlying ischemic bowel and thus added to the bowel's ischemia. He demonstrated that adhesions act as a vascular graft, stimulated by the presence of tissue ischemia. This present study suggests a similar role for the omentum. When placed in contact with an anastomotic gap, or length of ischemic gut, the omentum responds with an outgrowth of richly vascular tissue, which then acts as the major vascular source to the wound. This ability of the omentum to produce a neovasculature to ischemic tissues has been demonstrated in the kidney, 17brain, TM and spleen, ~9among other tissues. The same phenomenon has more recently been described for the induction of a blood supply to experimentally induced tumors in animals, g~ leading to a measurable increase in tumor size. In the present study, there was no detectable difference between a mechanically deficient and a profoundly ischemic anastomosis in their propensity to induce new vessels from the omentum. This may be because the ischemia produced by cutting the bowel and placing sutures provides a maximal ischemic stimulus in itself for neovasculature development. An alternative explanation may be the "angiogenic factor" as proposed by Cartier e t aL g~
Vol. 35, No. 8
EFFECT OF AN ANASTOMOTIC OMENTAL WRAP
This is a property possessed by the omental fat, which acts independently from ischemia, to induce vascular neogenesis. In Cartier e t al.'s study an omental lipid extract induced new vessels within a rabbit cornea to a greater extent than that produced by perirenal fat. Tisinai e t a L 22 have also found the omentum to more effectively produce new vessels than perirenal tissue. In their model a free graft of small intestine was placed within either the omenrum or the renal capsule of the fetal rat. A significantly greater number of intestines induced a blood supply and remained viable in the omental group than in the renal capsule group. In view of these findings, and the wealth of other properties the greater om ent um has been found to possess in humans, it appears to be an ideal structure for a protective role. It has been shown to absorb fluid and particulate matter 23 and transport them to tissue macrophages and immunocompetent c e l l s . 24' 25 It has been found to readily envelope injured tissues 23 and to control hemorrhage. 26 It would appear to be intrinsically better than adhesions from other organs for assisting injured or ischemic tissues. However, this ability was not evident in our study. In those rats surviving beyond 48 hours in which randomly occurring circumferential adhesions had developed, an abundance of granulation tissue was produced and, as for those with an omental wrap, a neovasculature was able to develop from the third day. These adhesions, predominantly composed of loops of small bowel, were thus producing a neovasculature as effectively as the omentum. Whether the small bowel contains the same angiogenic factor as omentum remains to be studied. The value of a viable omental wrap has clinical implications for those gastrointestinal anastomoses with a known propensity to leak, particularly extraperitoneal esophageal and low rectal joints. In both these settings successful clinical experience with an omental wrap has been reported. 27-29 Goligher e t al, 3~ reported in 1970 a very high rate of anastomotic dehiscence occurring after anterior resection, despite a meticulous technique. Overall, 51 percent of high and low anterior resections leaked, with the observation made that most leaks occurred posteriorly toward the bared presacral space. Of particular interest was the fact that 26 of the 37 leaks did not present with a fecal fistula or peritonitis but were locally contained on barium enema appearance. This implies that adhe-
737
sions to surrounding tissues were of clinical importance to contain these anastomotic leaks. This p h e n o m e n o n was examined in a very elegant study performed by McLachlin e t al. 31 in 1976. Using a dog model, they found that, if an extensive pelvic dissection left an ischemic rectal anastomosis without direct contact to surrounding soft tissues, there was a high incidence of major leaks. The local placement of vascularized omentum was able to reverse this effect and dramatically improve anastomotic competence. Goligher e t al.'s clinical findings and McLachlin e t al.'s experimental findings are consistent with the findings of this current study, as it has shown the avidity with which soft tissues will adhere to an anastomotic defect and contribute significantly to its healing, if placed in direct contact with it. The advantage of the omental wrap appears to be its guarantee of providing circumferential contact. CONCLUSION The omental pedicle wrap provides effective protection for a compromised intestinal anastomosis. It provides circumferential soft tissue contact, which can act as a plug to prevent early anastomotic leakage, and a source of granulation tissue and a neovasculature for later wound repair. REFERENCES 1. Bennett WH. A case of perforating gastric ulcer in which the opening, being otherwise intractable, was closed by means of an omental plug: recovery. Lancet 1896;2:310-1. 2. Waiters W. An omental flap in transperitoneal repair of recurring vesicovaginal fistulas. Surg Gynecol Obstet 1937;64:74-5. 3. Wilkie DP. Some functions and surgical uses of the omentum. BMJ 1911;2:1103-6. 4. Fabian TC, Stone HH. Arrest of severe liver hemorrhage by an omental pack. South Med J 1980; 73:1487-90. 5. Dupont C, Menard Y. Transposition of the greater omentum for reconstruction of the chest wall. Plast Reconstr Surg 1972;49:263-7. 6. Turner-Warwick R. The use of the omental pedicle graft in urinary tract reconstruction. J Urol 1976; 1116:341-7. 7. Goldsmith HS, Beattie EJ. Protection of vascular prostheses following radical inguinal excisions. Surg Clin North Am 1969;49:413-9. 8. Goldsmith HS. Protection of low rectal anastomosis with intact omentum. Surg Gynecol Obstet 1977;
738
ADAMS E T A L
144:584-6. 9. Lanter B, Mason R. Use of omental pedicle graft to protect low anterior colonic anastomosis. Dis Colon Rectum 1979;22:448-51. 10. Goldsmith HS, Kiely AA, Randall HT. Protection of intrathoracic oesophageal anastomoses by omentum. Surgery 1968;63:464-6. 11. McLachlin AD, Denton DW. Omental protection of omental anastomoses. Am J Surg 1972;125:134-40. 12. Katsikas D, Sechas M, Antypas G, et al. Beneficial effect of omental wrapping of unsafe intestinal anastomoses. Int Surg 1977;62;8:435-7. 13. Carter DC, Jenkins DHR, Whitfield HN. Omental reinforcement of intestinal anastomoses. Br J Surg 1972;59;2:129-32. 14. Gulati SM, Thusoo TK, Kakar A, Iyenger B, Pandey KK. Comparative study of free omental, peritoneal, dacron velour and marlex mesh reinforcement of large bowel anastomosis. Dis Colon Rectum 1982; 25;6:517-21. 15. Ravitch MM, Brolin R, Kolter J, Yap S. Studies in the healing of intestinal anastomoses. World J Surg 1981;5:627-37. 16. Ellis H. The aetiology of post-operative abdominal adhesions. Br J Surg 1962;50:10-6. 17. Eliska O. Blood flow in the revascularised kidney. Revascularisation by means of omentum and an iliac loop. J Cardiovasc Surg (Torino) 1968;9:342-7. 18. Goldsmith HS, Chen WF, Duckett S. Prevention of cerebral infarction in the dog by intact omentum. Am J Surg 1975;130:317-20. 19. De Renzi E, Boeri G. Isolated spleen tissue revascularised by the greater omentum. Ber Klin Wochenschr 1903;40:773-5. 20. Chalmers PJ, Newing RK. Influence ofomental trans-
21.
22.
23. 24. 25.
26.
27.
28.
29.
30.
31.
Dis Colon Rectum, August 1992 position on experimental tumours. J Surg Oncol 1986;32:135-7. Cartier R, Brunette I, Hashimoto K, Bourne WM, Schaff HV. Angiogenic factor: a possible mechanism for neovascularisation produced by omental pedicles. J Thorac Cardiovasc Surg 1990;99:264-8. Tisinai K, Shedd F, Harris R, et al. Comparison of growth, neovascularisation and enzymatic function of fetal intestinal grafts in the omentum and renal capsule. J Pediatr Surg 1990;25:914-6. Wilkie DPD. Some functions and surgical uses of the omentum. BMJ 1911;2:1103-6. Dudgeon IS, Ross A. Experiments on the great omenturn. Am J Med Sci 1906;132:37-41. Walker FC, Rogers AW. The greater omentum as a site of antibody synthesis. Br J Exp Pathol 1961; 42:222-31. Davis CB. Free transplantation of the omentum subcutaneously and within the abdomen. JAMA 1917; 68:705-6. Goldsmith HS, Kiely AA, Randall HT. Protection of intrathoracic oesophageal anastomoses by omentum. Surgery 1968;63:464-6. Lanter B, Mason RA. Use of omental pedicle graft to protect low anterior colonic anastomosis. Dis Colon Rectum 1979;22:448-51. Goldsmith HS. Protection of low rectal anastomosis with intact omentum. Surg Gynecol Obstet 1977; 144:584-6. Goligher JC, Graham NG, De Dombal FT. Anastomotic dehiscence after anterior resection of rectum and sigmoid. Br J Surg 1970;57:109-18. McLachlin AD, Olsson LS, Pitt DF. Anterior anastomosis of the rectosigmoid colon: an experimental study. Surgery 1976;80:306-11.
T
No reprints are available.
anastomoses in o m e n t u m to r e d u c e the risk of leakage. However, the experimental evidence to support this is contradictory, n 13 In this study we have used an animal m o d e l to investigate the effects of an omental wrap w h e n placed around a potentially leaking anastomosis.
MATERIALS AND METHODS Adult 300-g Wistar rats were used in the study. They were h o u s e d and bred in a vivarium in the grounds of the Westmead Hospital, u n d e r the care of veterinary staff e m p l o y e d by the University of Sydney. Approval to c o n d u c t this study was granted by the C u m b e r l a n d Area Health Service Research Committee. Three anastomotic techniques were investigated: a control anastomosis, w h i c h p r o d u c e d a reliable join, a mechanically deficient anastomosis, and an ischemic anastomosis. These latter two t e c h n i q u e s p r o d u c e d c o m p r o m i s e d anastomoses w h i c h simulated, in a s o m e w h a t exaggerated fashion, the anastomotic p r o b l e m s of mechanical failure and ischemia. The e x p e r i m e n t was then p e r f o r m e d in two parts. In the first part, the effect of an omental wrap on anastomotic c o m p e t e n c e was determined. The seco n d part of the e x p e r i m e n t investigated the histologic events occurring in the incorporation of o m e n t u m into a c o m p r o m i s e d anastomosis and the timing of the induction of a neovasculature.
Anastomotic Technique operations were p e r f o r m e d through
The a midline incision u n d e r intraperitoneal ketamine anesthesia, using a sterile technique. The animals were given access to only water and dilute lactulose solution for 24 hours preoperatively to clear the bowel and were elevated on a grid off the floor of their cage to prevent coprophagy. 731
ADAMS E T A L
732
The male rats have a large retractile testis with an associated pad of fat which is in fact larger than the omentum. This was found in early experiments to wrap itself around the anastomoses and so confuse the results. For this reason, this fat pad was routinely ligated and excised in all rats in the study. The mid small bowel was divided at the point of insertion of a mesenteric vessel and then reanastomosed. A continuous Connell suture with 7/0 silk was used for the control anastomosis. Two techniques were used to produce compromised anastomoses. The first was mechanically deficient in that only three interrupted sutures were used to oppose the bowel ends. The second was performed with four interrupted sutures and then made ischemic by ligating the mesenteric vessel and the marginal vessels 2 cm on either side of the anastomosis (Fig. 1). In the rat the omentum hangs free from the greater curve of the stomach and is not attached to the colon. To perform the omental wrap the anastomosed bowel was brought up to lie adjacent to the stomach. The omentum was then wrapped around the anastomosis to completely encircle it. It was maintained in position by sutures placed between the base of the omentum and the bowel
C.
Dis Colon Rectum, August 1992
mesentery behind the anastomosis and from the tip of the omentum to the bowel mesentery in front of the anastomosis (Fig. 2).
Determination of Anastomotic
Competence For the first part of the experiment, 50 rats underwent laparotomy and bowel anastomosis. Ten animals were randomly assigned to each of five experimental groups: a control group, a deficient anastomotic group, an ischemic anastomotic group, a deficient group with an omental wrap, and an ischemic group with an omental wrap. All rats that died underwent postmortem examination on the day of death. Survivors were sacrificed at six weeks. At postmortem examination the anastomoses were assessed for the presence of intraperitoneal leakage, perianastomotic abscess, and the formation of adhesions. The propensity to develop adhesions was assessed by measurement of the proportion of the circumference of the anastomosis adherent to surrounding organs.
a
be d.
Figure 1. Diagram of small bowel anastomoses, a. Bowel divided opposite mesenteric vessel, b. Control anastomosis; continuous suture with 7/0 silk. c. "Deficient anastomosis"; three interrupted sutures, d. "lschemic anastomosis"; four anastomotic sutures. The mesenteric vessels have been ligated 2 cm distant from either side of the anastomosis.
Figure 2. Diagram of omental wrap. a. Omentum sutured to bowel mesentery posteriorly, b. Circumferential wrap and omentum sutured anteriorly.
Vol. 35, No. 8
EFFECT OF AN ANASTOMOTIC OMENTAL WRAP
The anastomotic segment was then excised. It was fixed in formalin and prepared in a paraffin block. After staining with hematoxylin and eosin, it was submitted to histologic examination. Histologic Events For the second part of the experiment a sample of rats from each anastomotic group was sacrificed on selected postoperative days for submission to histologic study. The numbers sacrificed from each group, and the days of sacrifice, are shown in Table 1. Because the unwrapped anastomoses had a much higher mortality rate in the early postoperative period, it was found necessary to perform more of these anastomoses to produce sufficient numbers for later examination. To demonstrate the vascular contribution from the omentum to the healing anastomosis, a dye was injected into the omental vasculature at the time of the rat's sacrifice. This technique involved reanesthetizing the rat and exploring the abdomen. The anastomosed segment of bowel was isolated from the rest of the gut and its mesentery and was left attached only to the omental pedicle. The chest was then opened and the descending aorta cannulated. India ink was then injected until the omental Table 1. Timing of Postmortem Examination After Intestinal Anastomosis Number of Rats with each Anastomotic Technique
Day of Examination Ischemic Deficient Ischemic Deficient with with (no (no Wrap Wrap Wrap) Wrap) 1 2 3 4 5 7 10 14 18 21 42 Total
1" 2 2 2 2 2 2 1" 1" 1" 2 18
0 2 2 2 2 2 2 1 1" 0 2 16
12" 141" 2 1 0 0 0 1 0 0 0
9* 131" 2 1 1 0 1 0 0 0 0
30
27
All rats, except those marked with an asterisk, were sacrificed and underwent a dye study to outline the omental vasculature. * Spontaneous deaths, no dye injected. 1"Two sacrificed and dye study performed; remainder were spontaneous deaths.
733
vessels were seen to fill. The anastomosed segment of gut was then excised and examined. An analogous study was also performed on those rats sacrificed without an omental wrap in which spontaneous adhesions surrounded the anastomosis, the anastomosed segment of gut was isolated on a pedicle of adhesions, instead of omentum, and a similar injection technique was used. Under the microscope the dye could clearly be seen filling the omental vessels, including the capillaries. As capillaries budded into new vessels in the granulation tissue, these also filled with dye and their progress could be mapped. Statistical M e t h o d s Differences between anastomotic groups for both survival and the propensity to induce adhesions were compared using Fisher's exact test. The level for statistical significance was set at P < 0.05. RESULTS Survival In the first experiment there was a statistically significant survival advantage found with the addition of an omental wrap, in both deficient and ischemic anastomoses (Table 2). Fifteen of the 20 rats with either ischemic or deficient unwrapped anastomoses died in the first 48 hours. Inspection of the peritoneal cavity in each case revealed feculent peritonitis with obvious leaks from anastomotic gaps. Microscopic examination demonstrated infarcted bowel edges and an absence of bowel apposition. Three rats died after the first week as a result of bowel obstruction from anastomotic stricture. In each case a perianastomotic abscess was present, enclosed within surrounding adhesions. The addition of the omental wrap reduced the incidence of early fatalities but did not affect the incidence of delayed anastomotic stricture. There were three strictures in the deficient group with a wrap and two in the ischemic group with a wrap. In the deficient group, one stricture led to a fatal obstruction and two produced an incomplete obstruction in survivors at six weeks. In the rats with ischemic anastomoses and a wrap, two deaths followed early anastomotic leaks, one of which as a result of the omentum having been twisted during its preparation and rendered nonviable, and two deaths occurred later from obstruction.
ADAMS E T AL
734
Dis Colon Rectum, August 1992
Table 2. Number of Rats Surviving to Six Weeks Anastomotic Technique Used
Rats Dying Before Six Weeks
Rats Alive at Six Weeks
Total
Deficient Ischemic Deficient with omental wrap Ischemic with omental wrap Control
9* 10t 1" 4t 1
1* 0t 9* 6t 9
10 10 10 10 10
* P < 0.0005. t P < 0.005.
Table 3. Incidence of Anastomotic Adhesions (Measured as a Percentage of the Anastomotic Circumference Involved in Adhesions) Number of Rats with Anastomotic Adhesions
Anastomotic Technique Used
Deficient Ischemic Deficient with omental wrap Ischemic with omental wrap Control
No Adhesions
Adhesions Less Than 50% of Anastomosis
Adhesions More Than 50% of Anastomosis
0 0 7 6 0
0 0 2 2 0
10 10 1 2 9*
* One unable to be assessed.
There was one death in the control group. The cause of death was unclear since the anastomosis was identified within the remains and found to be intact. The other nine rats survived to six weeks with intact and patent joints. On microscopic examination these had healed by primary intention with inverted bowel edges and mucosal apposition.
Adhesions All anastomoses not wrapped in omentum developed adhesions to surrounding intraperitoneal structures. In all rats surviving beyond 48 hours, these were completely circumferential. In contrast to this, 13 of the 20 anastomoses wrapped with omentum formed no adhesions at all (Table 3). This difference between wrapped and unwrapped groups was statistically significant ( P < 0.001). In the seven rats with an omental wrap in which some perianastomotic adhesions were present, an inflammatory reaction extended through the full thickness of the overlying omentum. In the remaining 13 cases without adhesions, the inflammatory reaction was confined to the luminal surface of the omentum.
Patterns Observed in the Healing of Anastomoses with an Omental Wrap In the second part of the experiment, there were five deaths among the 34 wrapped anastomoses, one from an early leak and four from bowel obstruction. The ischemic preparation was found to consistently produce a 15- to 20-mm length of infarcted bowel at the anastomotic site. The deficient technique produced anastomotic gaps 3 to 5 mm in length. In the first 48 hours, the omentum became adherent to the ischemic bowel wall and bridged any anastomotic defects (Fig. 3). By day 3, granulation tissue was seen to develop on the luminal surface of the omentum. At this stage dye was seen to pass from omental to bowel wall vessels in all four cases studied (Figs. 4 and 5), demonstrating the induction of a vascular anastomosis between the omental blood supply and the anastomosed bowel. Over the next six weeks, the infarcted bowel edges were resorbed and collagen was laid down in the cylinder of omentum joining the separated ends of bowel. In those cases in which bowel obstruction
Vol. 35, No. 8
EFFECT OF AN ANASTOMOTIC OMENTAL WRAP
Figure 3. Ischemic anastomosis wrapped in omentum, at 48 hours. One side of the anastomosis is shown. The omentum can be seen spanning the gap between the infarcted bowel edges (marked by curved arrows). Acute inflammatory changes are present on the luminal surface of the omentum (shown by straight arrow). Hematoxylin and eosin, x29.
735
Figure 5. Higher power of Figure 4 to show dye within the bowel wall capillaries. Hematoxylin and eosin, x350.
Figure 6. Deficient anastomosis wrapped in omentum, at 14 days. There is maturation of the granulation tissue and opening up of vascular channels. Hematoxylin and eosin,
x26. Figure 4. Deficient anastomosis wrapped in omentum, at 72 hours. Granulation tissue has formed on the luminal surface of the omentum (orientated toward the top of the picture). Black dye can be seen within the bowel wall vessels (arrows), better seen on higher magnification in Figure 5. Hematoxylin and eosin, x40. developed, an exuberant granulation tissue response had arisen from the omentum and occluded the lumen. Vascular anastomoses persisted between the omentum and bowel wall (Fig. 6). At six weeks some mucosal proliferation became evident at the scar edges. Patterns Observed in the Healing of Anastomoses Without a n O m e n t a l W r a p Fifty-seven rats were examined after the construction of compromised anastomoses without the
addition of an omental wrap. Forty-four died in the first 48 hours. In each of these cases, obvious leaks were seen at postmortem examination, through anastomotic gaps not covered by adhesions. Thirteen rats with unwrapped anastomoses were sacrificed and examined between days 2 and 14. In each case fully circumferential adhesions were present at the anastomosis. The adhesions were formed predominantly from loops of small intestine. Often liver, pancreas, abdominal wall, and sometimes a small tongue of omentum also contributed to the adhesion complex. The pattern of healing in those rats surviving beyond 48 hours was similar to that seen in rats with an omental wrap. The adhesions acted as a plug to the anastomotic defects to prevent early leakage. The luminal surface of the adhesions then
ADAMS E T A L
736
developed an inflammatory reaction, followed by the production of the major portion of the granulation tissue in the wound. In the two rats examined on days 10 and 14, this granulation tissue had matured into a fibrous collagen plug. A vascular anastomosis between the adhesions and the bowel wall was demonstrated in three of the four rats examined on day 3 and in all rats examined subsequently. DISCUSSION Using a lethal anastomotic technique, the omental pedicle wrap has provided a dramatic improvement in experimental animal survival by maintaining intestinal continuity and integrity over a distance between two separated ends of bowel. The omental wrap forms an effective bridge over anastomotic defects in the first 48 hours and then provides the bulk of the granulation tissue in their subsequent healing. An anastomosis develops between the omental and bowel wall vessels as early as the third postoperative day to aid in anastomotic healing, but the input of omental blood arrives too late to avoid bowel edge infarction from the original ischemic injury. Random adhesions around a compromised anastomosis can behave in a similar fashion to an omental wrap, including the development of a neovasculature. However, these adhesions do not reliably form a seal in the first 48 hours. The survival advantage found in the present study is consistent with the experience of McLachlin and Denton ~ and Katsikas e t aL Ig who both tested the omental wrap on ischemic intestinal anastomoses in dogs, but contrasts with the experience of Carter and colleagues, 13 who tested both pedicled and free omental wraps on a much less lethal anastomotic model in the rabbit. The problem identified with wrapped anastomoses in these studies, as in our own, was the development of delayed anastomotic strictures and bowel obstruction. However, we have found that these strictures have resulted from early anastomotic failure With bowel separation and infarction. The stricture results from the profuse inflammatory response in the surrounding wrap, followed by an exuberant growth of granulation tissue. In these cases the absence of a wrap would have inevitably led to an early leak and peritonitis. The fact that so many animals with a wrapped compromised anas-
Dis Colon Rectum, August 1992
tomosis actually had an untroubled convalescence with an intact anastomosis at six weeks demonstmtes the wrap's protective ability. Another feature seen in association with a wrap was the reduced incidence of perianastomotic adhesions, not only in comparison with unwrapped compromised anastomoses, but in comparison with controls. This was found to be associated with the containment of the perianastomotic inflammation to the luminal surface of the omentum. The biologic viability of the wrap appears to be critical for its function. The one case in this series in which the wrap was accidentally devascularized led to its early failure. Devascularized omentum and other tissues have been previously studied for their potential to protect anastomoses. < 11, 13. 14 It has been found that the devascularized omentum is not of benefit and actually leads to a worse anastomotic outcome than no wrapping at all. This phenomenon was well demonstrated by Ravitch e t al. 15 when they induced a high leak rate from anastomoses isolated from surrounding adhesions by a wrap of polyethylene or glove rubber. Ellis le found that a free omental flap actually drew a blood supply away from underlying ischemic bowel and thus added to the bowel's ischemia. He demonstrated that adhesions act as a vascular graft, stimulated by the presence of tissue ischemia. This present study suggests a similar role for the omentum. When placed in contact with an anastomotic gap, or length of ischemic gut, the omentum responds with an outgrowth of richly vascular tissue, which then acts as the major vascular source to the wound. This ability of the omentum to produce a neovasculature to ischemic tissues has been demonstrated in the kidney, 17brain, TM and spleen, ~9among other tissues. The same phenomenon has more recently been described for the induction of a blood supply to experimentally induced tumors in animals, g~ leading to a measurable increase in tumor size. In the present study, there was no detectable difference between a mechanically deficient and a profoundly ischemic anastomosis in their propensity to induce new vessels from the omentum. This may be because the ischemia produced by cutting the bowel and placing sutures provides a maximal ischemic stimulus in itself for neovasculature development. An alternative explanation may be the "angiogenic factor" as proposed by Cartier e t aL g~
Vol. 35, No. 8
EFFECT OF AN ANASTOMOTIC OMENTAL WRAP
This is a property possessed by the omental fat, which acts independently from ischemia, to induce vascular neogenesis. In Cartier e t al.'s study an omental lipid extract induced new vessels within a rabbit cornea to a greater extent than that produced by perirenal fat. Tisinai e t a L 22 have also found the omentum to more effectively produce new vessels than perirenal tissue. In their model a free graft of small intestine was placed within either the omenrum or the renal capsule of the fetal rat. A significantly greater number of intestines induced a blood supply and remained viable in the omental group than in the renal capsule group. In view of these findings, and the wealth of other properties the greater om ent um has been found to possess in humans, it appears to be an ideal structure for a protective role. It has been shown to absorb fluid and particulate matter 23 and transport them to tissue macrophages and immunocompetent c e l l s . 24' 25 It has been found to readily envelope injured tissues 23 and to control hemorrhage. 26 It would appear to be intrinsically better than adhesions from other organs for assisting injured or ischemic tissues. However, this ability was not evident in our study. In those rats surviving beyond 48 hours in which randomly occurring circumferential adhesions had developed, an abundance of granulation tissue was produced and, as for those with an omental wrap, a neovasculature was able to develop from the third day. These adhesions, predominantly composed of loops of small bowel, were thus producing a neovasculature as effectively as the omentum. Whether the small bowel contains the same angiogenic factor as omentum remains to be studied. The value of a viable omental wrap has clinical implications for those gastrointestinal anastomoses with a known propensity to leak, particularly extraperitoneal esophageal and low rectal joints. In both these settings successful clinical experience with an omental wrap has been reported. 27-29 Goligher e t al, 3~ reported in 1970 a very high rate of anastomotic dehiscence occurring after anterior resection, despite a meticulous technique. Overall, 51 percent of high and low anterior resections leaked, with the observation made that most leaks occurred posteriorly toward the bared presacral space. Of particular interest was the fact that 26 of the 37 leaks did not present with a fecal fistula or peritonitis but were locally contained on barium enema appearance. This implies that adhe-
737
sions to surrounding tissues were of clinical importance to contain these anastomotic leaks. This p h e n o m e n o n was examined in a very elegant study performed by McLachlin e t al. 31 in 1976. Using a dog model, they found that, if an extensive pelvic dissection left an ischemic rectal anastomosis without direct contact to surrounding soft tissues, there was a high incidence of major leaks. The local placement of vascularized omentum was able to reverse this effect and dramatically improve anastomotic competence. Goligher e t al.'s clinical findings and McLachlin e t al.'s experimental findings are consistent with the findings of this current study, as it has shown the avidity with which soft tissues will adhere to an anastomotic defect and contribute significantly to its healing, if placed in direct contact with it. The advantage of the omental wrap appears to be its guarantee of providing circumferential contact. CONCLUSION The omental pedicle wrap provides effective protection for a compromised intestinal anastomosis. It provides circumferential soft tissue contact, which can act as a plug to prevent early anastomotic leakage, and a source of granulation tissue and a neovasculature for later wound repair. REFERENCES 1. Bennett WH. A case of perforating gastric ulcer in which the opening, being otherwise intractable, was closed by means of an omental plug: recovery. Lancet 1896;2:310-1. 2. Waiters W. An omental flap in transperitoneal repair of recurring vesicovaginal fistulas. Surg Gynecol Obstet 1937;64:74-5. 3. Wilkie DP. Some functions and surgical uses of the omentum. BMJ 1911;2:1103-6. 4. Fabian TC, Stone HH. Arrest of severe liver hemorrhage by an omental pack. South Med J 1980; 73:1487-90. 5. Dupont C, Menard Y. Transposition of the greater omentum for reconstruction of the chest wall. Plast Reconstr Surg 1972;49:263-7. 6. Turner-Warwick R. The use of the omental pedicle graft in urinary tract reconstruction. J Urol 1976; 1116:341-7. 7. Goldsmith HS, Beattie EJ. Protection of vascular prostheses following radical inguinal excisions. Surg Clin North Am 1969;49:413-9. 8. Goldsmith HS. Protection of low rectal anastomosis with intact omentum. Surg Gynecol Obstet 1977;
738
ADAMS E T A L
144:584-6. 9. Lanter B, Mason R. Use of omental pedicle graft to protect low anterior colonic anastomosis. Dis Colon Rectum 1979;22:448-51. 10. Goldsmith HS, Kiely AA, Randall HT. Protection of intrathoracic oesophageal anastomoses by omentum. Surgery 1968;63:464-6. 11. McLachlin AD, Denton DW. Omental protection of omental anastomoses. Am J Surg 1972;125:134-40. 12. Katsikas D, Sechas M, Antypas G, et al. Beneficial effect of omental wrapping of unsafe intestinal anastomoses. Int Surg 1977;62;8:435-7. 13. Carter DC, Jenkins DHR, Whitfield HN. Omental reinforcement of intestinal anastomoses. Br J Surg 1972;59;2:129-32. 14. Gulati SM, Thusoo TK, Kakar A, Iyenger B, Pandey KK. Comparative study of free omental, peritoneal, dacron velour and marlex mesh reinforcement of large bowel anastomosis. Dis Colon Rectum 1982; 25;6:517-21. 15. Ravitch MM, Brolin R, Kolter J, Yap S. Studies in the healing of intestinal anastomoses. World J Surg 1981;5:627-37. 16. Ellis H. The aetiology of post-operative abdominal adhesions. Br J Surg 1962;50:10-6. 17. Eliska O. Blood flow in the revascularised kidney. Revascularisation by means of omentum and an iliac loop. J Cardiovasc Surg (Torino) 1968;9:342-7. 18. Goldsmith HS, Chen WF, Duckett S. Prevention of cerebral infarction in the dog by intact omentum. Am J Surg 1975;130:317-20. 19. De Renzi E, Boeri G. Isolated spleen tissue revascularised by the greater omentum. Ber Klin Wochenschr 1903;40:773-5. 20. Chalmers PJ, Newing RK. Influence ofomental trans-
21.
22.
23. 24. 25.
26.
27.
28.
29.
30.
31.
Dis Colon Rectum, August 1992 position on experimental tumours. J Surg Oncol 1986;32:135-7. Cartier R, Brunette I, Hashimoto K, Bourne WM, Schaff HV. Angiogenic factor: a possible mechanism for neovascularisation produced by omental pedicles. J Thorac Cardiovasc Surg 1990;99:264-8. Tisinai K, Shedd F, Harris R, et al. Comparison of growth, neovascularisation and enzymatic function of fetal intestinal grafts in the omentum and renal capsule. J Pediatr Surg 1990;25:914-6. Wilkie DPD. Some functions and surgical uses of the omentum. BMJ 1911;2:1103-6. Dudgeon IS, Ross A. Experiments on the great omenturn. Am J Med Sci 1906;132:37-41. Walker FC, Rogers AW. The greater omentum as a site of antibody synthesis. Br J Exp Pathol 1961; 42:222-31. Davis CB. Free transplantation of the omentum subcutaneously and within the abdomen. JAMA 1917; 68:705-6. Goldsmith HS, Kiely AA, Randall HT. Protection of intrathoracic oesophageal anastomoses by omentum. Surgery 1968;63:464-6. Lanter B, Mason RA. Use of omental pedicle graft to protect low anterior colonic anastomosis. Dis Colon Rectum 1979;22:448-51. Goldsmith HS. Protection of low rectal anastomosis with intact omentum. Surg Gynecol Obstet 1977; 144:584-6. Goligher JC, Graham NG, De Dombal FT. Anastomotic dehiscence after anterior resection of rectum and sigmoid. Br J Surg 1970;57:109-18. McLachlin AD, Olsson LS, Pitt DF. Anterior anastomosis of the rectosigmoid colon: an experimental study. Surgery 1976;80:306-11.
