©1991 S. Kargcr AG, Basel OOI4-312X/9I/0232-OIOOS2.75/0
Eur Surg Res 1991;23:100-107
Effect of Fecal Loading with/without Peritonitis on the Healing of a Colonic Anastomosis: An Experimental Study B. Ravo3, N. Metwallb, J. Yehc, P. Polanskyd, F.M. Fratlarolia aII Surgical Clinic, University of Rome, Italy: bState University of New York, Stony Brook, N.Y., USA; cWinthrop University Hospital, Mineola, N.Y., USA; d McGraw-Hill, New York, N.Y., USA
Key Words. Intracolonic bypass procedure • Peritonitis • Anastomotic healing • Coloshield • Colorectal anastomosis
Classical surgical teaching cautions against primary anastomosis in the presence of peritonitis due to delayed anastomotic healing and high leak rate [1-4], Some of the
important factors responsible for the high leak rate in emergency surgery are: the un prepared colon, the edematous bowel wall and the inflamed peritoneum. However, pre
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
Abstract. An experimental randomized prospective study was carried out in 64 dogs to evaluate the effect of intraluminal fecal matter at the anastomosis with/without peritonitis on the healing of a colonic anastomosis. The animals, none of whom had bowel preparation, were randomized in four groups: group I sigmoid resection and anastomosis, group II sig moid resection and intraluminal fecal diversion from the anastomosis, group III induced fecal peritonitis, sigmoid resection and anastomosis and group IV induced fecal peritonitis, sigmoid resection and intraluminal fecal diversion from the anastomosis. Forty-eight hours before sacrifice at 5, 10 and 15 days, 10 pCi/kg C m proline was given intravenously. Speci mens were analyzed for hydroxyproline content, tissue counts and specific activity. The tissue counts and specific activity were analyzed by three-way analysis of variance. Overall, regardless of the groups, there was a statistically significant decrease in specific activity and tissue count from day 5 to day 15 and day 10 to day 15 at the anastomosis (p < 0.05). When comparing groups II and IV to groups I and III, there was a significant increase in specific activity and tissue count at the anastomosis of group II and IV (p < 0.05). This experimental study demonstrates that early anastomotic healing can occur even in presence of treated peritonitis as long as the fecal matter is diverted and prevented from coming in contact with the anastomotic site without disrupting the bowel continuity or function.
Healing of Colonic Anastomosis and Peritonitis
Materials and Methods Sixty-four adult female mongrel dogs were pro spectively randomized into four groups. Randomiza tion was by envelope on the day of surgery. Group I (14) underwent sigmoid resection and anastomosis. Group II (15) underwent sigmoid resection, anasto mosis and IBP. Group III (18) underwent the induc tion of fecal peritonitis, sigmoid resection and anas tomosis. Group IV (17) underwent the induction of fecal peritonitis, sigmoid resection, anastomosis and IBP. In group III and IV, the peritonitis was induced by means of a 2-cm-long colostomy on the antimesenteric border of a sigmoid colon through a small lower midline incision. The colon was thus returned to the peritoneal cavity and the abdominal incision closed. Therapy for the induced peritonitis com menced after 3 h of peritonitis. Therapy included intravenous fluid resuscitation, antibiotic adminis tration and a laparotomy. Laparotomy included peri toneal toilette and excision of the injured colon with primary anastomosis. All animals were cared for by the same animal handlers and surgeons, did not receive bowel preparation, were treated by the same diet, intravenous solutions, antibiotics and solid food started on the 3rd postoperative day. Anesthesia con sisted of nitrous oxide, oxigen and halothane. The animals in group I and II received four doses of intra venous sodium cefoxitan and the animals in group III and IV received intravenous chloramphenicol and sodium penicillin G for 10 days. The anastomoses were performed with 12 interrupted 3-0 silk sutures. In groups III and IV 3 liters of saline solution were used for intraoperative peritoneal lavage. The ani mals were sacrificed at 5, 10 and 15 days. Forty-eight hours before sacrifice, 10pCi/kg proline was given intravenously to the surviving animals. All anasto moses were evaluated at the time of sacrifice as reported in previous communication [16]. All speci mens of animals that were injected with C ,4 proline were analyzed for hydroxyproline content and ra dioactivity counts. Quantitation of hydroxyproline content and ra dioactivity was done just after sacrifice of the ani mals according to the method of Peterofsky and Prockop [17], Forty-eight hours prior to sacrifice. Cu proline (268.8 pCi/mmol, New England Nuclear, Boston. Mass.) was injected intraveneously into each dog in a dose of 10 pCi/kg of body weight. At sacri fice 1 cm of colon above and below the anastomosis
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
vious investigations studying the effect of peritonitis on anastomotic healing have studied the effect of peritonitis on the heal ing of an anastomosis often with anasto motic dehiscences and leakage and therefore continuous peritonitis even after the perito nitis was treated [5-13]. With the introduc tion of the intracolonic bypass procedure (IBP) a model has been provided to study the effect of peritonitis on the healing of a co lonic anastomosis without continuous peri tonitis formation. The IBP diverts the fecal flow and intestinal secretions from the intra luminal anastomotic site. It thereby creates a temporary intraluminal barrier between the bowel wall and the fecal matter, excluding the anastomosis from the fecal flow and pos sible anastomotic leakage, even in the pres ence of dehiscences, as demonstrated in pre vious experimental studies [14, 15], For a better understanding it should be pointed out that although the terms dehiscences and leakage are used interchangeably in the liter ature, the meanings are different. A dehis cence is a separation of an anastomosis and a leakage is the escape of the intraluminal fecal content outside the bowel wall. Therefore, when a dehiscence occurs, there may not necessarily be a leak such as when the dehis cence is plugged by omentum or a contigu ous organ, ot when the intraluminal fecal content is diverted by a colostomy or the IBP. However, when an anastomosis leaks, it is always preceded by a dehiscence and gives clinical complications. To better understand the role of intralu minal fecal matter and extraluminal perito neal inflammation on dehiscence and possi ble leakage formation and healing of a co lonic anastomosis, an experimental random ized prospective study has been carried out.
101
102
[ 19].
Table 1. Mean and standard errors of specific ac tivity of hydroxyproline (10'' cpm/mg) for each indi vidual and total groups and days at the anastomosis Group
Number of dogs
Day
Mean
SDT error of mean
I
6 4 3
5 10 15
5.2 5.0 4.6
0.47 0.49 0.49
II
6 4 5
5 10 15
6.8 7.1 4.0
0.80 0.24 0.42
III
6 3 4
5 10 15
6.4 4.8 4.0
0.54 0.55 0.33
IV
6 3 3
5 10 15
6.6 6.5 6.6
0.37 0.85 0.59
24 14 15
5 10 15
6.26* 5.89 4.94
0.56 0.53 0.40
I-IV
*
p > 0.05.
Results Of the 64 dogs, 53 dogs survived and were injected with C m proline before sacrifice (ta ble 1). One dog died in group I of an abdom inal wound rupture, 4 dogs died in group III and 4 in group IV because of sepsis. Two dogs, 1 in group III and 1 in group IV. were not injected because of technical situations. The clinical results have been reported pre viously [16]. Results indicated a significant statistical increase in leak rate, p < 0.02, when comparing the combined 25% leak rate of group I and III (anastomosis with intraluminal fecal contact) to the combined 0% leak rate of group II and IV (anastomosis excluded from intraluminal fecal contact) re-
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
and 2 cm of right colon were removed in each dog, coded and sent to the laboratory in a blind fashion for hydroxyproline extraction and radioactivity counts. The colonic tissue was then washed with cold saline solution. Both segments were weighed and chopped into fine pieces before homogenization by a blender and polytron (Brinkman, West Gurie, N.Y.). The homogenated solution was placed in dialysis sacs (molecolar weight cutoff was 12,000) and dialyzed in 20 vol of 0.15 mol of saline at 4°C for 24 h. The nondialyzable material was then harvested and an aliquot was hydrolyzed in 6 N hydrochloric acid at 100°C for 48 h. An aliquot of the nondialyzable solution was also used to be mixed with Scinti verse (Fishers, Springfield, N.J.) and to count the radioac tivity by liquid scintillation counter (LKB1210). Por tions of the hydrolysate were then analyzed for hy droxyproline by modification of the method of Prockop and Udenfriend [18]. The hydrolysate was neutralyzed to pH 0.8 ± 0.5 with 6 N and 1 N KOH. Phenolphthalein was used as an indicator. After satu ration with K.CI, 5 ml of 10% alanine (pH 8.7) and 1 ml of 1 mol potassium borate buffer (pH 8.7) were added to the sample, then mixed well on a vortex. The mixture was oxidized at room temperature with 1 ml of 2 M chloramine T solution. After 25 min the oxidation was stopped by adding 3 ml of 3.6 M so dium thisulfatc solution.The solution was extracted with 5 ml of toluene by vigorous shaking and the acqueous phase was then chapped and heated in a boiling water bath for 30 min to form pyrrole; after cooling the lube, 5 ml toluene was added again and the sample was shaken vigorously. The sample was centrifuged briefly at low speed and 2.5 ml of toluene extract was pipetted to a clean test tube to react with Ehrlich’s reagent. The resulting color was red after 30 min at 560 pm on Beckman Model 35 spectropho tometer. Tissue count-total collagen radioactivity was mea sured (cpm/g tissue) of the nondialyzable tissue ho mogenate with a molecular weight of 12,000. Specific activity (cpm/g OH proline) newly synthesized colla gen calculated dividing the tissue count (cpm/g tissue) by the hydroxyproline content (cpm/mg tissue). The statistic studies were carried out by three-way analysis of variance. The independent factors being day (5. 10 and 15), peritonitis and the use of the IBP in preventing fecal contact at the anastomosis. The technique of the IBP has been previously described
Ravo/Mctwall/Yeh/Polansky/Frattaroli
103
Fig. 1. Specific activities (SA) of left colon anasto mosis of group I (anastomosis with intraluminal fecal contact) and group II (anastomosis without intralumi nal fecal contact).
Fig. 2. Specific activities (SA) of left colon anasto mosis of group III (anastomosis with intraluminal fecal contact and peritonitis) and group IV (anasto mosis without intraluminal fecal contact but with peritonitis).
gardless of peritonitis. The tissue counts and specific activity analyzed by three-way anal ysis of variance demonstrated negligible in teraction between the presence of peritonitis and the use of bypass procedure meaning that the effect of protecting the anastomosis from fecal contact by the IBP was consistent regardless of the presence of peritonitis. The interaction of the above with day was also negligible. Therefore, the final model in cluded only three main effects: days, perito nitis and intraluminal fecal contact. F tests were performed using the residual mean square error to test among levels of the main effects (table 1). Overall, regardless of the groups, there was a statistically significant decrease in spe-
cific activity and tissue count from day 5 to day 15 and day 10 to day 15 (p < 0.05) at the anastomosis (table 1). But at the right colon segments there was no decrease in specific activity or tissue counts. When the peritoni tis group III and IV are compared to the non peritonitis groups I and II, there is an in crease in specific activity at the anastomosis which is not statistically significant. When comparing the anastomosis of groups II and IV to groups I and III (fig. 1,2, table 2), there is a significant increase in specific activity and tissue count at the anastomosis of groups II and IV (p < 0.05). When compar ing tissue counts of the right colon segments of the same groups, there are not statistically significant differences. When comparing
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
Healing of Colonic Anastomosis and Peritonitis
104
Ravo/Metwall/Yeh/Polansky/Frattaroli
Table 2. Means and statistically significant differences of specific activity of hydroxyproline (10’ cpni/mg) at the anastomosis with peritonitis vs. nonperitonitis and fecal exclusion from the anastomosis with IBP vs. fecal contact at the anastomosis Anastomosis with
Group
Number of dogs
Mean
SDT error of mean
Peritonitis Nonperitonitis
III, IV I, II
25 28
5.93 5.66
0.53 0.63
Fecal exclusion (IBP) Fecal contact
II, IV I. Ill
27 26
6.40* 5.15
0.54 0.44
*
p < 0.05.
Table 3. Means and statistically significant differences of specific activity of hydroxyproline (10’ cpm/mg) of the right colon with peritonitis vs. nonperitonitis and fecal exclusion from the anastomosis with IBP vs. fecal contact at the anastomosis Right colon with
Group
Number of dogs
Means
SDT error of mean
Peritonitis Nonperitonitis
III, IV I, II
25 28
4.07 4.15
0.46 0.50
Fecal exclusion (IBP) Fecal contact
II, IV I, III
27 26
4.41* 3.79
0.61 0.36
p < 0.05.
specific activity of the right segment, there is an increase in group II and IV over group I and III, which is of statistical significance (p < 0.05) (table 3).
Discussion The intestinal tract is like a tunnel ‘an elongated passageway open at both ends’, therefore it communicates with the outside environment making the lumen a contami nated dirty area and the intraluminal gas trointestinal tract itself external to the body like the skin. The intestinal wall is the natu
ral barrier between the contaminated intesti nal lumen and the clean sterile peritoneal, retroperitoneal or perintestinal spaces. Most of the complications in gastrointes tinal surgery arise when the intestinal wall is violated by the pathological process, foreign bodies, etc., or when a resection is carried out and an anastomosis is performed. Over the years surgeons have tried to overcome the danger of anastomotic leakage, in elec tive surgery, by cleaning the intestinal lumen well before surgery, correct possible meta bolic causes that affect the healing of an anastomosis and make sure that anastomosis is well vascularized and free of tension. In
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
*
emergency situations the anastomosis is usually delayed, a colostomy is created and at a later date an elective anastomosis is car ried out. With the IBP a temporary artificial intestinal wall is created that diverts the fecal matter intraluminally from the anasto mosis, therefore protecting the anastomosis and permitting primary anastomosis in one operation even in an emergency situation. A colonic anastomosis heals in about 1214 days. The first 3-5 days is the most criti cal period in the healing of a colonic anasto mosis because the two approximate cut ends of the bowel are held together mainly by the uniting sutures or staples [5, 10, 20, 21 ]. Goligher et al. [22] have demonstrated by gastrografin enema after low anterior resection a 69% leak rate at the anastomosis site proba bly corresponding to the dehiscence forma tion and not leakage. The fact that not a large number of anastomotic dehiscences leak clinically is most likely because these dehis cences are plugged by omentum or contigu ous organs. MacLaughlin and Olson [23] have shown a statistically significant de crease in anastomotic leak rate in low ante rior anastomosis when the dead space about the anastomosis is filled with a pedical graft of viable omentum rather than when blood alone or blood and Escherichia coli are about pelvic anastomosis; when the omentum or contiguous organs are prevented to come in contact with an anastomosis by a polyethyl ene or glove rubber, there is a 100% leakage as reported by Ravitch et al. [ 11 ]; most likely anastomotic gaps between sutures result in dehiscence from the time of operation espe cially in the presence of peritonitis where there is an increased leak which delays anas tomotic healing as shown from previously reported studies [5, 13] and group III in this study. This experimental study demon
105
strates that anastomosis leakage is the cause of continuous peritonitis in group III be cause in group IV where a temporary barrier was created by the coloshield between the anastamosis and the intraluminal fecal mat ter there were no leakages and there was also early healing as indicated by increased hydroxyproline. The method we used to mea sure hydroxyproline involves toluene extrac tion to remove proline prior to the boiling procedure to convert hydroxyproline to pyr role form derivative. From the results of a preliminary study, we measured the radioac tivity in the toluene extract after the boiling procedure (presumably all the pyrrole form derivative of hydroxyproline). We found that 92.5 ± 4.6% (mean ± SD) of the cpm radioactivity measured in the nondialyzable solution were recovered in this fraction. This result indicates that 92.5 ± 4.6% of the radioactivity measured in the nondialyzable solution were of hydroxyproline origin, which existed only in collagen. It appears from experimental data that fecal loading (maximal intraluminal fecal matter) is a very important factor in leakage formation, per sistent peritonitis and delayed healing as in group I and III of this study and as also dem onstrated by Smith [24] and others in the past [5, 13]. In group II and IV were the fecal loading was prevented from coming in contact with the anastomosis by the IBP. there was a sta tistically significant increase in specific ac tivity of hydroxyproline at the anastomosis even in the presence of peritonitis as in group IV. Because fecal loading carries a high incidence of anastomotic leakage in co lonic surgery [2, 22, 24, 25], these experi mental results may explain why antimicro bial preparation of the bowel and adequate bowel preparation or intraoperative lavage
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
Healing of Colonic Anastomosis and Peritonitis
of the unprepared colon may be beneficial to anastomotic healing [2, 24, 26 27], The in creased specific activity of the right colon in group II and IV might be due to the IBP itself or feces in the remaining colon as Jiborn et al. [20, 21] have also described when there is colonic fecal stasis. There are advantages of having an anastomosis pro tected intraluminally from the feces without preventing the fecal flow from proximal to distal colon; the action of the fecal content on the colon may be important in stimulat ing collagen and protein synthesis by stretch ing the bowel wall [28-31 ], when fecal flow is diverted as with temporary colostomy there is decreased collagen in the excluded colonic segment [31]. In conclusion, this experimental study demonstrates that early anastomotic healing can occur even in the presence of treated peritonitis and possible other situations where the colon is unprepared or inflamed as long as the intraluminal fecal content is di verted and prevented from coming in con tact with the anastomotic site without dis rupting the bowel continuity or function. Clinical results have corroborated the exper imental ones [19, 32, 33].
References 1 Fielding LP, Stewart Brown S, Biesoovsky L, Kearney G: Anastomotic integrity after operation for large bowel cancer, a multicentre study. Br Med J 1980;281:411-414. 2 Goligher JC: Surgery of the Anus, Rectum and Colon, cd 4. London. Baillière, Tindall, 1980. 3 Killingback M: Management of perforated diverticulits. Surg Clin North Am 1983;63:97-115. 4 Madden JL, Tan PV: Primary resection and anas tomosis in the treatment of perforated lesion of the colon with abscess or diffuse perilonitis. Am J Surg 1961;1 13:646-650.
Ravo/Melwall/Yeh/Polansky/Frattaroli
5 Cronin K, Jackson DS. Dunphy JE: Changing bursting strength and collagen content of the heal ing colon. Surg Gynecol Obstet 1968; 126:747— 750. 6 Cronin K, Jackson DS, Dunphy JE: Specific activ ity of hydroxyproline tritum in the healing colon. Surg Gynecol Obstet 1968; 126:1061 -1065. 7 Dunphy JE: The cat gut. Am J Surg 1971 ; 119:1 — 8. 8 Hawley PR. Page-Faulk W: A circulatory collagé nase inhibitor. Br J Surg 1970;57:900-904. 9 Irvin TT: Collagen metabolism in infected colonic anastomosis. Surg Gynecol Obstet 1976; 143:220— 224. 10 Irvin TT: Wound Healing Principle and Practice. London, Chapman-Hall, 1981. 11 Ravitch MM, Brolin R, Kolter J, et al: Studies in the healing of intestinal anastomosis. World J Surg 1981;5:627-637. 12 Smith M, Enquest IF: A quantitative study of impaired healing resulting from infection. Surg Gynecol Obstet 1967;125:965-973. 13 Yamakawa T, Patin S, Sobels, Morgenstern L: Healing of colonic anastomoses following resec tion for experimental diverticulitis. Arch Surg 1971;109:17-20. 14 Ravo B, Ger R: Intracolonic by-pass by an intra luminal tube; an experimental study. Dis Colon Rectum 1884;27:360-366. 15 Ravo B, Ger R: Management of esophageal dehis cence by an intraluminal by-pass tube; an experi mental study. Am J Surg 1985;149:733-738. 16 Ravo B, Metwaily N, Castera P, el al: The impor tance of intraluminal anastomotic fecal contact and peritonitis in colonic anastomotic leakages; an experimental study. Dis Colon Rectum 1988; 31:868-871. 17 Peterofsky B, Prockop DJ: A method for the simultaneous measurement of the radioactivity of proline C m and hydroxyproline C M in biological materials. Anal Biochem 1962;4:400-406. 18 Prockop DJ, Udcnfriend S: A specific method for the analysis of hydroxyproline in tissue and urine. Anal Biochem 1960;1:228-229. 19 Ravo B: How I do it: The intracolonic by-pass procedure. Int J Colored Dis 1987;2:38-42. 20 Jiborn H, Ahonen J, Zederfeldt B: Healing of experimental colonic anastomoses: The effect of suture technique on collagen concentration on the colonic wall. Am J Surg 1978;135:333-340.
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
106
107
Healing of Colonic Anastomosis and Peritonitis
29 Freihoffer U, Robinson EA. Jackson DS. Dunphy JE: The effect of bacterial endotoxin on connec tive tissue growth and wound tensile strength. Surg Forum 1960;11:293-295. 30 Palmer RM, Reeds PJ, Lobley GE, Smith RH: The effect of intermittent changes in tension on protein and collagen synthesis in isolated rabbit muscles. Biochem J 1981;198:491-498. 31 Blomquist P. Jiborn H, Zederfeldt B: Effect of diverting colostomy on collagen metabolism in the colonic wall. Am J Surg 1985;149:330-333. 32 Keane PF, Ohri SK, Wood CB, Sackier JM: Man agement of the obstructed left colon by the onestage ¡ntracolonic by-pass procedure. Dis Colon Rectum 1988;31:948-951. 33 Ravo B, Mishrick A, Addei K. et al: The treatment of perforated diverticulitis by one-stage intraco lonic by-pass procedure. Surgery 1987; 102:771 — 776. Received: June 11, 1990 Accepted: January 28, 1991 Biagio Ravo, MD Rome American Hospital Via Emilio Longoni 69 1-00155 Rome (Italy)
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
21 Jiborn H. Ahonen J. Zederfeldt B: Healing of experimental colonic anastomoses: The effect of suture technique on collagen metabolism in the colonic wall. Am J Surg 1980;139:406-413. 22 Goligher JC, Graham NG. De Dombal FT: Anas tomotic dehiscence after anterior resection of rec tum and sigmoid. Br J Surg 1970;57:109-118. 23 MacLaughlin AD, Olson LJ: Anastomosis of the rectosigmoid colon; an experimental study. Sur gery 1976;80:306-311. 24 Smith SRG, Connely JC, Gilmore OJA: The effect of fecal loading on colonic anastomotic healing. Br J Surg 1983;70:49-50. 25 Irvin TT, Goligher JC: Aetiology of diruption of intestinal anastomosis. Br J Surg 1973;60:461— 464. 26 Moore EE, Shannon FI, Moore FA, McCroskey BL: Value of distal colon washout in civilian rectal trauma. Reducing gut bacterial translocation. J Trauma 1988;28:989-992. 27 Foster ME, Johnson CD, Billing PF, et al: Intra operative antegrade lavage and anastomosis heal ing in acute colonic obstruction. Dis Colon Rec tum 1986;29:255-259. 28 Wahl SM: Role of mononuclear cells in the wound repair process; in Dineen P (ed): Surgical Wound. Philadelphia. Lea & Febiger, 1981, pp 63-75.
Eur Surg Res 1991;23:100-107
Effect of Fecal Loading with/without Peritonitis on the Healing of a Colonic Anastomosis: An Experimental Study B. Ravo3, N. Metwallb, J. Yehc, P. Polanskyd, F.M. Fratlarolia aII Surgical Clinic, University of Rome, Italy: bState University of New York, Stony Brook, N.Y., USA; cWinthrop University Hospital, Mineola, N.Y., USA; d McGraw-Hill, New York, N.Y., USA
Key Words. Intracolonic bypass procedure • Peritonitis • Anastomotic healing • Coloshield • Colorectal anastomosis
Classical surgical teaching cautions against primary anastomosis in the presence of peritonitis due to delayed anastomotic healing and high leak rate [1-4], Some of the
important factors responsible for the high leak rate in emergency surgery are: the un prepared colon, the edematous bowel wall and the inflamed peritoneum. However, pre
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
Abstract. An experimental randomized prospective study was carried out in 64 dogs to evaluate the effect of intraluminal fecal matter at the anastomosis with/without peritonitis on the healing of a colonic anastomosis. The animals, none of whom had bowel preparation, were randomized in four groups: group I sigmoid resection and anastomosis, group II sig moid resection and intraluminal fecal diversion from the anastomosis, group III induced fecal peritonitis, sigmoid resection and anastomosis and group IV induced fecal peritonitis, sigmoid resection and intraluminal fecal diversion from the anastomosis. Forty-eight hours before sacrifice at 5, 10 and 15 days, 10 pCi/kg C m proline was given intravenously. Speci mens were analyzed for hydroxyproline content, tissue counts and specific activity. The tissue counts and specific activity were analyzed by three-way analysis of variance. Overall, regardless of the groups, there was a statistically significant decrease in specific activity and tissue count from day 5 to day 15 and day 10 to day 15 at the anastomosis (p < 0.05). When comparing groups II and IV to groups I and III, there was a significant increase in specific activity and tissue count at the anastomosis of group II and IV (p < 0.05). This experimental study demonstrates that early anastomotic healing can occur even in presence of treated peritonitis as long as the fecal matter is diverted and prevented from coming in contact with the anastomotic site without disrupting the bowel continuity or function.
Healing of Colonic Anastomosis and Peritonitis
Materials and Methods Sixty-four adult female mongrel dogs were pro spectively randomized into four groups. Randomiza tion was by envelope on the day of surgery. Group I (14) underwent sigmoid resection and anastomosis. Group II (15) underwent sigmoid resection, anasto mosis and IBP. Group III (18) underwent the induc tion of fecal peritonitis, sigmoid resection and anas tomosis. Group IV (17) underwent the induction of fecal peritonitis, sigmoid resection, anastomosis and IBP. In group III and IV, the peritonitis was induced by means of a 2-cm-long colostomy on the antimesenteric border of a sigmoid colon through a small lower midline incision. The colon was thus returned to the peritoneal cavity and the abdominal incision closed. Therapy for the induced peritonitis com menced after 3 h of peritonitis. Therapy included intravenous fluid resuscitation, antibiotic adminis tration and a laparotomy. Laparotomy included peri toneal toilette and excision of the injured colon with primary anastomosis. All animals were cared for by the same animal handlers and surgeons, did not receive bowel preparation, were treated by the same diet, intravenous solutions, antibiotics and solid food started on the 3rd postoperative day. Anesthesia con sisted of nitrous oxide, oxigen and halothane. The animals in group I and II received four doses of intra venous sodium cefoxitan and the animals in group III and IV received intravenous chloramphenicol and sodium penicillin G for 10 days. The anastomoses were performed with 12 interrupted 3-0 silk sutures. In groups III and IV 3 liters of saline solution were used for intraoperative peritoneal lavage. The ani mals were sacrificed at 5, 10 and 15 days. Forty-eight hours before sacrifice, 10pCi/kg proline was given intravenously to the surviving animals. All anasto moses were evaluated at the time of sacrifice as reported in previous communication [16]. All speci mens of animals that were injected with C ,4 proline were analyzed for hydroxyproline content and ra dioactivity counts. Quantitation of hydroxyproline content and ra dioactivity was done just after sacrifice of the ani mals according to the method of Peterofsky and Prockop [17], Forty-eight hours prior to sacrifice. Cu proline (268.8 pCi/mmol, New England Nuclear, Boston. Mass.) was injected intraveneously into each dog in a dose of 10 pCi/kg of body weight. At sacri fice 1 cm of colon above and below the anastomosis
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
vious investigations studying the effect of peritonitis on anastomotic healing have studied the effect of peritonitis on the heal ing of an anastomosis often with anasto motic dehiscences and leakage and therefore continuous peritonitis even after the perito nitis was treated [5-13]. With the introduc tion of the intracolonic bypass procedure (IBP) a model has been provided to study the effect of peritonitis on the healing of a co lonic anastomosis without continuous peri tonitis formation. The IBP diverts the fecal flow and intestinal secretions from the intra luminal anastomotic site. It thereby creates a temporary intraluminal barrier between the bowel wall and the fecal matter, excluding the anastomosis from the fecal flow and pos sible anastomotic leakage, even in the pres ence of dehiscences, as demonstrated in pre vious experimental studies [14, 15], For a better understanding it should be pointed out that although the terms dehiscences and leakage are used interchangeably in the liter ature, the meanings are different. A dehis cence is a separation of an anastomosis and a leakage is the escape of the intraluminal fecal content outside the bowel wall. Therefore, when a dehiscence occurs, there may not necessarily be a leak such as when the dehis cence is plugged by omentum or a contigu ous organ, ot when the intraluminal fecal content is diverted by a colostomy or the IBP. However, when an anastomosis leaks, it is always preceded by a dehiscence and gives clinical complications. To better understand the role of intralu minal fecal matter and extraluminal perito neal inflammation on dehiscence and possi ble leakage formation and healing of a co lonic anastomosis, an experimental random ized prospective study has been carried out.
101
102
[ 19].
Table 1. Mean and standard errors of specific ac tivity of hydroxyproline (10'' cpm/mg) for each indi vidual and total groups and days at the anastomosis Group
Number of dogs
Day
Mean
SDT error of mean
I
6 4 3
5 10 15
5.2 5.0 4.6
0.47 0.49 0.49
II
6 4 5
5 10 15
6.8 7.1 4.0
0.80 0.24 0.42
III
6 3 4
5 10 15
6.4 4.8 4.0
0.54 0.55 0.33
IV
6 3 3
5 10 15
6.6 6.5 6.6
0.37 0.85 0.59
24 14 15
5 10 15
6.26* 5.89 4.94
0.56 0.53 0.40
I-IV
*
p > 0.05.
Results Of the 64 dogs, 53 dogs survived and were injected with C m proline before sacrifice (ta ble 1). One dog died in group I of an abdom inal wound rupture, 4 dogs died in group III and 4 in group IV because of sepsis. Two dogs, 1 in group III and 1 in group IV. were not injected because of technical situations. The clinical results have been reported pre viously [16]. Results indicated a significant statistical increase in leak rate, p < 0.02, when comparing the combined 25% leak rate of group I and III (anastomosis with intraluminal fecal contact) to the combined 0% leak rate of group II and IV (anastomosis excluded from intraluminal fecal contact) re-
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
and 2 cm of right colon were removed in each dog, coded and sent to the laboratory in a blind fashion for hydroxyproline extraction and radioactivity counts. The colonic tissue was then washed with cold saline solution. Both segments were weighed and chopped into fine pieces before homogenization by a blender and polytron (Brinkman, West Gurie, N.Y.). The homogenated solution was placed in dialysis sacs (molecolar weight cutoff was 12,000) and dialyzed in 20 vol of 0.15 mol of saline at 4°C for 24 h. The nondialyzable material was then harvested and an aliquot was hydrolyzed in 6 N hydrochloric acid at 100°C for 48 h. An aliquot of the nondialyzable solution was also used to be mixed with Scinti verse (Fishers, Springfield, N.J.) and to count the radioac tivity by liquid scintillation counter (LKB1210). Por tions of the hydrolysate were then analyzed for hy droxyproline by modification of the method of Prockop and Udenfriend [18]. The hydrolysate was neutralyzed to pH 0.8 ± 0.5 with 6 N and 1 N KOH. Phenolphthalein was used as an indicator. After satu ration with K.CI, 5 ml of 10% alanine (pH 8.7) and 1 ml of 1 mol potassium borate buffer (pH 8.7) were added to the sample, then mixed well on a vortex. The mixture was oxidized at room temperature with 1 ml of 2 M chloramine T solution. After 25 min the oxidation was stopped by adding 3 ml of 3.6 M so dium thisulfatc solution.The solution was extracted with 5 ml of toluene by vigorous shaking and the acqueous phase was then chapped and heated in a boiling water bath for 30 min to form pyrrole; after cooling the lube, 5 ml toluene was added again and the sample was shaken vigorously. The sample was centrifuged briefly at low speed and 2.5 ml of toluene extract was pipetted to a clean test tube to react with Ehrlich’s reagent. The resulting color was red after 30 min at 560 pm on Beckman Model 35 spectropho tometer. Tissue count-total collagen radioactivity was mea sured (cpm/g tissue) of the nondialyzable tissue ho mogenate with a molecular weight of 12,000. Specific activity (cpm/g OH proline) newly synthesized colla gen calculated dividing the tissue count (cpm/g tissue) by the hydroxyproline content (cpm/mg tissue). The statistic studies were carried out by three-way analysis of variance. The independent factors being day (5. 10 and 15), peritonitis and the use of the IBP in preventing fecal contact at the anastomosis. The technique of the IBP has been previously described
Ravo/Mctwall/Yeh/Polansky/Frattaroli
103
Fig. 1. Specific activities (SA) of left colon anasto mosis of group I (anastomosis with intraluminal fecal contact) and group II (anastomosis without intralumi nal fecal contact).
Fig. 2. Specific activities (SA) of left colon anasto mosis of group III (anastomosis with intraluminal fecal contact and peritonitis) and group IV (anasto mosis without intraluminal fecal contact but with peritonitis).
gardless of peritonitis. The tissue counts and specific activity analyzed by three-way anal ysis of variance demonstrated negligible in teraction between the presence of peritonitis and the use of bypass procedure meaning that the effect of protecting the anastomosis from fecal contact by the IBP was consistent regardless of the presence of peritonitis. The interaction of the above with day was also negligible. Therefore, the final model in cluded only three main effects: days, perito nitis and intraluminal fecal contact. F tests were performed using the residual mean square error to test among levels of the main effects (table 1). Overall, regardless of the groups, there was a statistically significant decrease in spe-
cific activity and tissue count from day 5 to day 15 and day 10 to day 15 (p < 0.05) at the anastomosis (table 1). But at the right colon segments there was no decrease in specific activity or tissue counts. When the peritoni tis group III and IV are compared to the non peritonitis groups I and II, there is an in crease in specific activity at the anastomosis which is not statistically significant. When comparing the anastomosis of groups II and IV to groups I and III (fig. 1,2, table 2), there is a significant increase in specific activity and tissue count at the anastomosis of groups II and IV (p < 0.05). When compar ing tissue counts of the right colon segments of the same groups, there are not statistically significant differences. When comparing
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
Healing of Colonic Anastomosis and Peritonitis
104
Ravo/Metwall/Yeh/Polansky/Frattaroli
Table 2. Means and statistically significant differences of specific activity of hydroxyproline (10’ cpni/mg) at the anastomosis with peritonitis vs. nonperitonitis and fecal exclusion from the anastomosis with IBP vs. fecal contact at the anastomosis Anastomosis with
Group
Number of dogs
Mean
SDT error of mean
Peritonitis Nonperitonitis
III, IV I, II
25 28
5.93 5.66
0.53 0.63
Fecal exclusion (IBP) Fecal contact
II, IV I. Ill
27 26
6.40* 5.15
0.54 0.44
*
p < 0.05.
Table 3. Means and statistically significant differences of specific activity of hydroxyproline (10’ cpm/mg) of the right colon with peritonitis vs. nonperitonitis and fecal exclusion from the anastomosis with IBP vs. fecal contact at the anastomosis Right colon with
Group
Number of dogs
Means
SDT error of mean
Peritonitis Nonperitonitis
III, IV I, II
25 28
4.07 4.15
0.46 0.50
Fecal exclusion (IBP) Fecal contact
II, IV I, III
27 26
4.41* 3.79
0.61 0.36
p < 0.05.
specific activity of the right segment, there is an increase in group II and IV over group I and III, which is of statistical significance (p < 0.05) (table 3).
Discussion The intestinal tract is like a tunnel ‘an elongated passageway open at both ends’, therefore it communicates with the outside environment making the lumen a contami nated dirty area and the intraluminal gas trointestinal tract itself external to the body like the skin. The intestinal wall is the natu
ral barrier between the contaminated intesti nal lumen and the clean sterile peritoneal, retroperitoneal or perintestinal spaces. Most of the complications in gastrointes tinal surgery arise when the intestinal wall is violated by the pathological process, foreign bodies, etc., or when a resection is carried out and an anastomosis is performed. Over the years surgeons have tried to overcome the danger of anastomotic leakage, in elec tive surgery, by cleaning the intestinal lumen well before surgery, correct possible meta bolic causes that affect the healing of an anastomosis and make sure that anastomosis is well vascularized and free of tension. In
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
*
emergency situations the anastomosis is usually delayed, a colostomy is created and at a later date an elective anastomosis is car ried out. With the IBP a temporary artificial intestinal wall is created that diverts the fecal matter intraluminally from the anasto mosis, therefore protecting the anastomosis and permitting primary anastomosis in one operation even in an emergency situation. A colonic anastomosis heals in about 1214 days. The first 3-5 days is the most criti cal period in the healing of a colonic anasto mosis because the two approximate cut ends of the bowel are held together mainly by the uniting sutures or staples [5, 10, 20, 21 ]. Goligher et al. [22] have demonstrated by gastrografin enema after low anterior resection a 69% leak rate at the anastomosis site proba bly corresponding to the dehiscence forma tion and not leakage. The fact that not a large number of anastomotic dehiscences leak clinically is most likely because these dehis cences are plugged by omentum or contigu ous organs. MacLaughlin and Olson [23] have shown a statistically significant de crease in anastomotic leak rate in low ante rior anastomosis when the dead space about the anastomosis is filled with a pedical graft of viable omentum rather than when blood alone or blood and Escherichia coli are about pelvic anastomosis; when the omentum or contiguous organs are prevented to come in contact with an anastomosis by a polyethyl ene or glove rubber, there is a 100% leakage as reported by Ravitch et al. [ 11 ]; most likely anastomotic gaps between sutures result in dehiscence from the time of operation espe cially in the presence of peritonitis where there is an increased leak which delays anas tomotic healing as shown from previously reported studies [5, 13] and group III in this study. This experimental study demon
105
strates that anastomosis leakage is the cause of continuous peritonitis in group III be cause in group IV where a temporary barrier was created by the coloshield between the anastamosis and the intraluminal fecal mat ter there were no leakages and there was also early healing as indicated by increased hydroxyproline. The method we used to mea sure hydroxyproline involves toluene extrac tion to remove proline prior to the boiling procedure to convert hydroxyproline to pyr role form derivative. From the results of a preliminary study, we measured the radioac tivity in the toluene extract after the boiling procedure (presumably all the pyrrole form derivative of hydroxyproline). We found that 92.5 ± 4.6% (mean ± SD) of the cpm radioactivity measured in the nondialyzable solution were recovered in this fraction. This result indicates that 92.5 ± 4.6% of the radioactivity measured in the nondialyzable solution were of hydroxyproline origin, which existed only in collagen. It appears from experimental data that fecal loading (maximal intraluminal fecal matter) is a very important factor in leakage formation, per sistent peritonitis and delayed healing as in group I and III of this study and as also dem onstrated by Smith [24] and others in the past [5, 13]. In group II and IV were the fecal loading was prevented from coming in contact with the anastomosis by the IBP. there was a sta tistically significant increase in specific ac tivity of hydroxyproline at the anastomosis even in the presence of peritonitis as in group IV. Because fecal loading carries a high incidence of anastomotic leakage in co lonic surgery [2, 22, 24, 25], these experi mental results may explain why antimicro bial preparation of the bowel and adequate bowel preparation or intraoperative lavage
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
Healing of Colonic Anastomosis and Peritonitis
of the unprepared colon may be beneficial to anastomotic healing [2, 24, 26 27], The in creased specific activity of the right colon in group II and IV might be due to the IBP itself or feces in the remaining colon as Jiborn et al. [20, 21] have also described when there is colonic fecal stasis. There are advantages of having an anastomosis pro tected intraluminally from the feces without preventing the fecal flow from proximal to distal colon; the action of the fecal content on the colon may be important in stimulat ing collagen and protein synthesis by stretch ing the bowel wall [28-31 ], when fecal flow is diverted as with temporary colostomy there is decreased collagen in the excluded colonic segment [31]. In conclusion, this experimental study demonstrates that early anastomotic healing can occur even in the presence of treated peritonitis and possible other situations where the colon is unprepared or inflamed as long as the intraluminal fecal content is di verted and prevented from coming in con tact with the anastomotic site without dis rupting the bowel continuity or function. Clinical results have corroborated the exper imental ones [19, 32, 33].
References 1 Fielding LP, Stewart Brown S, Biesoovsky L, Kearney G: Anastomotic integrity after operation for large bowel cancer, a multicentre study. Br Med J 1980;281:411-414. 2 Goligher JC: Surgery of the Anus, Rectum and Colon, cd 4. London. Baillière, Tindall, 1980. 3 Killingback M: Management of perforated diverticulits. Surg Clin North Am 1983;63:97-115. 4 Madden JL, Tan PV: Primary resection and anas tomosis in the treatment of perforated lesion of the colon with abscess or diffuse perilonitis. Am J Surg 1961;1 13:646-650.
Ravo/Melwall/Yeh/Polansky/Frattaroli
5 Cronin K, Jackson DS. Dunphy JE: Changing bursting strength and collagen content of the heal ing colon. Surg Gynecol Obstet 1968; 126:747— 750. 6 Cronin K, Jackson DS, Dunphy JE: Specific activ ity of hydroxyproline tritum in the healing colon. Surg Gynecol Obstet 1968; 126:1061 -1065. 7 Dunphy JE: The cat gut. Am J Surg 1971 ; 119:1 — 8. 8 Hawley PR. Page-Faulk W: A circulatory collagé nase inhibitor. Br J Surg 1970;57:900-904. 9 Irvin TT: Collagen metabolism in infected colonic anastomosis. Surg Gynecol Obstet 1976; 143:220— 224. 10 Irvin TT: Wound Healing Principle and Practice. London, Chapman-Hall, 1981. 11 Ravitch MM, Brolin R, Kolter J, et al: Studies in the healing of intestinal anastomosis. World J Surg 1981;5:627-637. 12 Smith M, Enquest IF: A quantitative study of impaired healing resulting from infection. Surg Gynecol Obstet 1967;125:965-973. 13 Yamakawa T, Patin S, Sobels, Morgenstern L: Healing of colonic anastomoses following resec tion for experimental diverticulitis. Arch Surg 1971;109:17-20. 14 Ravo B, Ger R: Intracolonic by-pass by an intra luminal tube; an experimental study. Dis Colon Rectum 1884;27:360-366. 15 Ravo B, Ger R: Management of esophageal dehis cence by an intraluminal by-pass tube; an experi mental study. Am J Surg 1985;149:733-738. 16 Ravo B, Metwaily N, Castera P, el al: The impor tance of intraluminal anastomotic fecal contact and peritonitis in colonic anastomotic leakages; an experimental study. Dis Colon Rectum 1988; 31:868-871. 17 Peterofsky B, Prockop DJ: A method for the simultaneous measurement of the radioactivity of proline C m and hydroxyproline C M in biological materials. Anal Biochem 1962;4:400-406. 18 Prockop DJ, Udcnfriend S: A specific method for the analysis of hydroxyproline in tissue and urine. Anal Biochem 1960;1:228-229. 19 Ravo B: How I do it: The intracolonic by-pass procedure. Int J Colored Dis 1987;2:38-42. 20 Jiborn H, Ahonen J, Zederfeldt B: Healing of experimental colonic anastomoses: The effect of suture technique on collagen concentration on the colonic wall. Am J Surg 1978;135:333-340.
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
106
107
Healing of Colonic Anastomosis and Peritonitis
29 Freihoffer U, Robinson EA. Jackson DS. Dunphy JE: The effect of bacterial endotoxin on connec tive tissue growth and wound tensile strength. Surg Forum 1960;11:293-295. 30 Palmer RM, Reeds PJ, Lobley GE, Smith RH: The effect of intermittent changes in tension on protein and collagen synthesis in isolated rabbit muscles. Biochem J 1981;198:491-498. 31 Blomquist P. Jiborn H, Zederfeldt B: Effect of diverting colostomy on collagen metabolism in the colonic wall. Am J Surg 1985;149:330-333. 32 Keane PF, Ohri SK, Wood CB, Sackier JM: Man agement of the obstructed left colon by the onestage ¡ntracolonic by-pass procedure. Dis Colon Rectum 1988;31:948-951. 33 Ravo B, Mishrick A, Addei K. et al: The treatment of perforated diverticulitis by one-stage intraco lonic by-pass procedure. Surgery 1987; 102:771 — 776. Received: June 11, 1990 Accepted: January 28, 1991 Biagio Ravo, MD Rome American Hospital Via Emilio Longoni 69 1-00155 Rome (Italy)
Downloaded by: Univ.of California Berkeley 128.32.10.230 - 8/3/2018 9:54:02 PM
21 Jiborn H. Ahonen J. Zederfeldt B: Healing of experimental colonic anastomoses: The effect of suture technique on collagen metabolism in the colonic wall. Am J Surg 1980;139:406-413. 22 Goligher JC, Graham NG. De Dombal FT: Anas tomotic dehiscence after anterior resection of rec tum and sigmoid. Br J Surg 1970;57:109-118. 23 MacLaughlin AD, Olson LJ: Anastomosis of the rectosigmoid colon; an experimental study. Sur gery 1976;80:306-311. 24 Smith SRG, Connely JC, Gilmore OJA: The effect of fecal loading on colonic anastomotic healing. Br J Surg 1983;70:49-50. 25 Irvin TT, Goligher JC: Aetiology of diruption of intestinal anastomosis. Br J Surg 1973;60:461— 464. 26 Moore EE, Shannon FI, Moore FA, McCroskey BL: Value of distal colon washout in civilian rectal trauma. Reducing gut bacterial translocation. J Trauma 1988;28:989-992. 27 Foster ME, Johnson CD, Billing PF, et al: Intra operative antegrade lavage and anastomosis heal ing in acute colonic obstruction. Dis Colon Rec tum 1986;29:255-259. 28 Wahl SM: Role of mononuclear cells in the wound repair process; in Dineen P (ed): Surgical Wound. Philadelphia. Lea & Febiger, 1981, pp 63-75.
