Neostigmine is frequently used to reverse the action of muscle relaxants after surgery and it has sometimes been used to treat paralytic post-operative ileus. However, since 1968, its use has been controversial after bowel anastomosis [1-11]. Many authors have suggested that neostigmine increases the rate of anastomotic disruption because of its prokinetic effect , and they have advised caution in its administration[1,7,8,10,11]. Moreover, other experimental studies have shown that the administration of other prokinetic agents, such as metoclopramide, during the early post-operative period is associated with an increase in the dehiscence of colonic anastomoses .
Other experiments in animals and clinical observations have indicated that neostigmine has no deleterious effect on bowel anastomosis [3-6].
There have been no studies in which accurate measures have been used to define the resistance of anastomoses after neostigmine treatment compared with resistance without such treatment. The bursting pressure and the bursting wall tension have been proposed as good indicators of anastomotic resistance [12-16], but these parameters have not been applied in attempts to assess the effects of neostigmine.
We postulated that stimulation of intestinal motility by a single dose of neostigmine after surgery (a usual final anaesthetic manoeuvre) would be insufficient to compromise the normal healing of a colonic anastomosis. The present report describes the results of a prospective randomized study in an animal model (the rat) in which different variables that have previously been used as indicators of anastomotic resistance were used to examine the effect of neostigmine on the colonic anastomosis.
Adult Sprague-Dawley rats of either sex (250-350 g body weight) were randomly assigned to each of two experimental groups, as follows, until a total of 20 surviving rats was obtained for each group on the fourth post-operative day: (1) control group, colonic anastomosis plus placebo after surgery (n=20); and (2) neostigmine group, colonic anastomosis plus a dose of neostigmine after surgery (n=20).
During the experiment, the animals were housed according to the European Community Standards for Care of Laboratory Animals (86/609/EEC). Standard rat chow (Panlab, Barcelona, Spain) and water were supplied ad libitum before surgery.
Animals were not deprived of food prior to surgery. The colon was not prepared before the operation and no preparations were made prior to anaesthesia. The anaesthetic protocol included ketamine (75 mg kg−1 body weight), xylazine (10 mg kg−1 body weight) and atropine (2 mg kg−1 body weight), all of which were administered intraperitoneally.
A 3-cm midline laparotomy was performed, equidistant from the pubis and the xiphoid process. The transverse colon was exteriorized and sectioned halfway between the hepatic and splenic angles, with care being taken to avoid damage to the marginal vascularization. A single-plane (all coats) end-to-end anastomosis of the extremities was then immediately made with eight non-absorbable 6/0 monofilament stitches. Finally, the abdominal wall was closed with a single-layer of continuous 1/0 silk sutures. All anastomoses were performed by the same surgeon.
After closure of the abdominal wall, all animals received an intravenous injection in a tail vein. In group 1, animals received 2 mL kg−1 body weight of saline solution and animals in group 2 received 1 mg kg−1 body weight of neostigmine(Prostigmine®, Roche, Madrid, Spain). All animals received the same volume of fluid. The dose of neostigmine was chosen on the basis of the results of Kishibayashi and colleagues , and a preliminary trial in four animals. These four animals received 2 mg kg−1 body weight of neostigmine after anastomosis and all of them died after administration of the drug, probably from the side effects of an overdose. Out of the rats in the study that received 1 mg kg−1 body weight of neostigmine, five animals died. Thus, we were clearly using a relatively high dose of the drug.
One treatment (placebo or neostigmine) was randomly assigned to each animal, and prior to the end of the experiment, the team that subsequently made the various measurements did not know which rats had received the drug.
During the post-operative period, animals were deprived of food but water was supplied ad libitum. Water consumption was recorded on a daily basis.
All animals were sacrificed on the fourth day after surgery by intracardiac injection of a lethal dose of thiopentone sodium under ketamine anaesthesia. With the animal in asystole, the abdomen was opened through two perpendicular (longitudinal and transverse) incisions along its entire length. The margins were separated by reference to four specific points, so that an excellent view of the peritoneal cavity was obtained.
Various records were then made, as follows:
- Water consumption.
- Clinical dehiscence. Dehiscence was evaluated in terms of the presence or absence of colonic contents within the peritoneal space.
- Size of the caecum. The caecum was measured twice with calipers: once after the anastomosis had been made; and then before removal of the surgical specimen on the day 4. The difference between the two values was calculated.
- Pre- and post-anastomotic colonic diameter. The diameters were measured by the procedure described above.
- Dilatation. Dilatation was considered present when the diameter of the pre-anastomotic transverse colon was at least 50% higher than the diameter of the post-anastomotic transverse colon.
- Obstruction. Obstruction was considered to exist when, in addition to dilatation as defined above, no faecal contents were present in the colon distal to the anastomosis. Obstruction may have been common on the fourth post-operative day as a consequence of the absence of intense peristalsis and because of marked inflammation of the anastomotic line.
- General adhesions. Adhesions were quantified according to the following scale: (0) no adhesions; (1) highly localized adhesions; (2) regional adhesions; and (3) diffuse adhesions.
- Adhesions to the anastomotic line (macroscopic assessment). The following protocol was employed to evaluate adhesions to the anastomotic line. The colon segment containing the anastomosis was removed by transverse sectioning of the colon at a distance of no less than 2 cm from either side of the anastomosis. At the same time and far from the anastomotic site, all structures with adhesions to the anastomotic line were sectioned to avoid detachment, and thus, to allow extraction of the specimen together with all local adhesions that might have developed. Having removed the segment of the colon, we used the following scale to define the proportion of the anastomotic line that was covered with adhesions: (1) 0-25% of the anastomotic line was covered with adhesions by adjacent structures; (2) 25-50%; (3) more than 75% but less than 100%; and (4) 100% of the entire circumference was covered with adhesions.
- Bursting pressure (BP) The surgical specimen was placed on a piece of white gauze in a bacteriological culture dish. The distal end of the colon segment was firmly ligated with 1/0 silk, while the proximal end was ligated to a 20-G intravenous perfusion catheter (Terumo®), with care being taken not to occlude the lumen. The catheter was linked, in turn, to a T-tube. One arm of the T-tube was connected to the capsule of a pressure transducer (no. 2028, Druck Ltd). The corresponding pressure signals were transmitted to a digital polygraph (MT6-PX, Lectromed). The other arm of the T-tube was connected to a perfusion pump (Braun) that had been loaded with a solution of methylene blue in distilled water. The entire pressure-recording system, up to the point where it was connected to the test specimen, contained distilled water[12-14](Fig. 1). The recording system was calibrated and set to zero immediately after connection of the specimen to the infusion circuit. Infusion was initiated at an infusion rate of 60 mL h−1, and was interrupted as soon as leakage of methylene blue was observed. Leakage was easily identified by the rapid staining of the gauze as a result of capillary action and by a sudden drop in pressure. The changes in pressure within the specimen were analysed with specific software (DS8C Program, Panlab, Barcelona, Spain). The BP was taken as the pressure at the moment when leakage was first noted at the anastomotic line.
- Radius of the anastomosis. Following measurement of the BP, the surgical specimen was fixed in formalin. No less than 4 days later, a scalpel was used to section the anastomosis, and the intraluminal diameter was measured with calipers, with care being taken to generate, manually, a circumference that was as close to a circle as possible. This method was adopted to avoid the problems associated with the considerable elasticity of the freshly dissected specimen, which could yield false values. The diameter recorded in this way differed from that of the non-fixed colon, but since this approach was used in all cases, we consider the results for different samples to be comparable.
- Bursting wall tension (BWT). The BWT is a measure of the circular tension of the bowel wall at the instant of rupture. It is a function of BP and radius, as reflected by Laplace's law. Thus, BWT=1.33 × 103 ×BP × R, where R is the intraluminal radius of the colon at the anastomotic line. The values obtained are expressed in dynes cm−1. Many authors consider the BWT to be the best indicator of anastomotic resistance .
We used Student's t-test and the χ2-test in the case of samples with a normal distribution, and the Mann-Whitney and Kruskal-Wallis tests for samples with a non-normal distribution.
Statistical significance was recognized at P<0.05.
Five rats that received neostigmine died after the injection, probably as a result of side effects, and these animals were excluded from the statistical analysis of results.
The change in body weight was the same in both groups, and no clinical dehiscence was recognized. The results obtained on the fourth post-operative day with respect to water consumption and variations in the size of the caecum are shown in Table 1. Table 1 shows that the rate of water consumption was higher in animals that had been treated with neostigmine (P=0.049). In all groups, the caecum on the fourth post-operative day had decreased in size, but the decrease was significantly greater in group 2 (P=0.029).
As shown in Fig. 2, dilatation was significantly more frequent in group 1 (nine cases) than in group 2 (one case) (P=0.01). Moreover, we found five obstructed anastomoses in group 1, while there were none in group 2 (P=0.047).
No statistical differences between the two groups were found in the case of changes in colonic diameters (pre- and post-anastomosis), general adhesions and adhesions to the anastomotic line.
Bursting pressure was similar in both groups: in group 1 it was 70.66±22.95 mmHg (mean±SD); and in group 2 it was 71.04±27.45 mmHg (mean±SD) (Fig. 2). In all cases, leakage of dye was seen at the anastomotic line. The mean radii and bursting wall tension were also similar in both groups.
The use of neostigmine to reverse the action of the muscle relaxant as a final anaesthetic manoeuvre in abdominal surgery has been widely discussed, particularly when a bowel anastomosis is performed [1-11]. It appears clear that the prokinetic effect of neostigmine occurs in both the small and large bowel; however, different studies have shown contradictory results about the intensity of this effect on each intestinal segment [1,3,11]. We chose colonic anastomosis as the experimental technique because colonic dehiscences are usually more frequent than small bowel dehiscences.
The risk of dehiscence caused by the drug's strong prokinetic effect is the reason that many surgeons do not like to administer neostigmine to their patients. For example, Bell and colleagues  reported a marked increase in the rate of anastomotic leakage when neostigmine was employed. They recorded the leakage rate three weeks after surgery by routine barium enema, but as they indicated, they did not know when leakage had happened. They suggested that it might have occurred in the early post-operative period as a result of the effect of neostigmine and excessive tension at the suture line. It is possible that other interfering factors, such as age and health status, were involved. By contrast, Cofer and colleagues  reported a retrospective study in which no anastomotic leakage was found in a group of patients who had received neostigmine to reverse the effects of d-tubocurarine. They evaluated anastomotic leakage by accurate clinical criteria. Many other authors have discussed the post-operative use of neostigmine in the context of experimental studies or in reviews of anaesthesia [2,3,5-11].
However, no studies involving accurate assessment of possible changes in anastomotic resistance have been reported. In our study, BP and BWT were measured on the fourth post-operative day, the last of the so-called 'critical days' for healing of an anastomosis [12-14,16,18], when its resistance is lowest. Many authors have suggested that BP or BWT are good indicators of anastomotic resistance [12-16]. We also recorded details of other parameters, such as water consumption (dehydration), body weight (nutritional depletion), clinical dehiscence, dilatation, obstruction, general adhesions (harmful adhesions) and adhesions to the anastomotic line (beneficial adhesions). All of these parameters are indicators of healing of an anastomosis, especially on the 'critical days' [12-14].
Bursting pressure and BWT were similar in both our groups (control and neostigmine) of animals. That is to say, resistance did not change when neostigmine was administered. This result was corroborated by the results for general adhesions and adhesions to the anastomotic line, which were similar in both groups. According to previous studies, these parameters are closely related to anastomotic bursting strength [12,13].
Our results do not contradict the opinion of many authors who have suggested that an increase in colonic motility during the early post-operative period might be deleterious. An earlier experimental study demonstrated the loss of anastomotic resistance when metoclopramide (a prokinetic agent) was administered to rats for 4 days after surgery, as a humane treatment for paralytic post-operative ileus . A single dose might be insufficient to cause injury to the anastomosis, as we observed with neostigmine treatment. We chose a high dose of neostigmine for our experiments and this dose had no deleterious effect, so it appears clear that the usual doses would also have no deleterious effect. Nevertheless, long-term treatment with neostigmine after an anastomosis might have the same effect as long-term treatment with metoclopramide.
We found that colonic dilatation and obstruction were more frequent in group 1 (control) than in group 2 (neostigmine). It is possible that neostigmine might prevent obstructions caused by peristaltic weakness because of its strong contractile effect on colonic smooth muscle which might induce an elimination of faeces.
To conclude, our results suggest that a single high dose of neostigmine has no deleterious effect on the resistance of colonic anastomoses. In our study, the colon was not prepared before surgery, which supports this conclusion. Although these unequivocal results in the laboratory require evaluation in a clinical setting, it has been shown that the inclusion of neostigmine in anaesthetic protocols should not decrease the resistance of colonic anastomoses or compromise normal healing. Moreover, obstructions caused by peristaltic weakness might be prevented by the elimination of faeces as a result of strong contraction of the colonic smooth muscle.
The authors thank Miss Carmen López-Sánchez and Mrs Trinidad Flores-Sánchez, Administrative Team, Education and Research Service, Albacete General Hospital, Albacete, Spain, for their invaluable help and generosity during the course of this study, and Dr J. Payá for his decisive help at the beginning of this study.
This work was supported in part by a grant from 'Fondo de Investigaciones Sanitarias' (Spain).
1 Bell CM, Lewis CB. Effect of neostigmine on integrity of ileorectal anastomoses. Brit Med J
2Brown EN, Daughety MJ, Petty WC. Integrity of intestinal anastomoses following muscle relaxant reversal with neostigmine. Anesth Analg
3 Yellin AE, Newman J, Donovan AJ. Neostigmine-induced hyperperistalsis. Effects on security of colonic anastomoses. Arch Surg
4 Cofer TW, Jr, Ray JE, Gathright JB. Does neostigmine cause disruption of large-intestinal anastomoses? A negative answer. Dis Colon Rectum
5 Morisot P, Loygue J, Guilmet C. [Effects of postoperative decurarization with neostigmine on digestive anastomoses.] Can Anaesth Soc J
6 Brolin RE, Ravitch MM. Studies in intestinal healing VI. Effect of pharmacological induced peristalsis on fresh intestinal anastomoses in dogs. Arch Surg
7 Buzello W, Krieg N, Brobmann GF. [Neostigmine and dehiscence of intestinal anastomoses.] Anasth Intensivther Notf Med
8 Aitkenhead AR. Anaesthesia for bowel surgery. Ann Chir Gynaecol
9 Herz BL. Colonic anastomotic disruption in myasthenia gravis. Report of two cases. Dis Colon Rectum
10 Carlstedt A, Nordgren S, Fasth S, Appelgren L, Hultén L. Epidural anaesthesia and postoperative colorectal motility. A possible hazard to a colorectal anastomosis. Int J Colorectal Dis
11 Ogilvy AJ, Smith G. The gastrointestinal tract after anaesthesia. Eur J Anaesthesiol
1995; 12 (Suppl. 10):
12 García-Olmo D, Payá J, Lucas FJ, García-Olmo DC. Effects of the pharmacological manipulation of postoperative intestinal motility on colonic anastomosis. An experimental study in a rat model. Int J Colorectal Dis
13 García-Olmo D, Lucas FJ, Payá J, López-Fando J. Relationship between peritoneal adhesion phenomena and the experimental resistance of colonic anastomoses: influence of omentoplasty. Eur Surg Res
14 García-Olmo D, Lucas FJ, Payá J. Effects of an intrabdominal latex (drain-like) tube on anastomotic resistance and formation of adhesions in a rat model. Dis Col Rectum
15 Adams W, Ctercteko G, Bilous M. Effect of an omental wrap on the healing and vascularity of compromised intestinal anastomoses. Dis Colon Rectum
16 Slavin J, Kingsnorth AN. Influence of methylprednisolone on the healing of intestinal anastomoses in rats. Br J Surg
17Kishibayashi N, Karasawa A. Stimulating effects of KW-5092, a novel gastroprokinetic agent, on the gastric emptying, small intestinal propulsion and colonic propulsion in rats. Jpn J Pharmacol
18Wise L, McAlister W, Stein T, Schuck P. Studies on the healing of anastomoses of small and large intestines. Surg Gynecol Obstet