5
Original article The interrupted serosubmucosal anastomosis – still the gold standard A. Leslie* and R. J. C. Steele* *Department of Surgery and Molecular Oncology, University of Dundee, Dundee, UK Received 8 July 2002; accepted 19 December 2002 Abstract Background The single-layer appositional serosub- mucosal anastomosis is a well established technique and appears to have a favourable record. Over a 15-year period the senior author of this paper has performed or directly supervised 553 anastomoses using this technique. This report describes the results of these operations, the results of stapled anastomoses carried out during the same period and discusses the utility of the handsewn technique. Methods From August 1986 to July 2001, 553 intes- tinal anastomoses in 550 patients were fashioned using single-layer, interrupted serosubmucosal 3 0 braided polyamide and 131 anastomoses in 131 patients were performed using a circular anastomosing stapler. Results One anastomotic leakage occurred in the group of patients whose anastomosis was handsewn (0.2%) and 11 leaks occurred in those who had a stapled anastomoses (8.4%). The mortality rate in each group was similar (2% and 2.3%, respectively). There were no deaths attributable to anastomotic dehiscence in either group. Conclusion In this prospectively audited series of 553 handsewn anastomoses the leakage rate was 0.2%. These results compare favourably with other published series and continue to support a single layer of interrupted serosubmucosal sutures as the gold standard for anasto- moses involving the large or small bowel. Keywords Serosubmucosal anastomosis, stapled anasto- mosis, technique Introduction The concept of using the submucosal layer of the bowel to hold stitches for a sound anastomosis was first introduced by William Halsted in his publication of 1887 [1], but this described work in dogs, and there is no evidence that Halsted ever applied his findings to man. It fell to Matheson in Aberdeen, working from the 1960s to the 1990s, to develop the technique for clinical use [2], and the extramucosal, appositional, interrupted sero- submucosal anastomosis has become widely used, with reported leakage rates in the region of 2% [3,4]. Since 1986 the senior author of this paper has employed this technique for all anastomoses involving the large and small bowel with the exceptions of upper gastrointestinal surgery, where a continuous sero- submucosal technique was used [5], and low rectal anastomoses, where stapling devices were employed. This report describes the results of a continuous series of operations carried out between August 1986 and July 2001 in four different centres either by, or under the direct supervision of, the senior author with a view to assessing the utility of the interrupted serosubmucosal anastomosis. It represents an update of a previous report [6], but on this occasion we have included all stapled low anastomoses carried out by the same surgeon in order to provide a complete picture of anastomosis-related mor- bidity. It should be emphasized, however, that the indications for sutured and stapled anastomoses in this series were quite separate in that the stapled approach was reserved for the lower anterior resections. This report therefore should not be seen as a direct comparison of the two techniques. Patients and methods Patients Over a 15-year period 553 anastomoses involving the large or small bowel in 550 patients (252 male, 298 Correspondence to: Ms A Leslie, Department of Surgery and Molecular Oncology, University of Dundee, Ninewells Hospital, Dundee, DD1 9SY, UK. E-mail: [email protected] 362 Ó 2003 Blackwell Publishing Ltd. Colorectal Disease, 5, 362–366

The interrupted serosubmucosal anastomosis – still the gold standard

Embed Size (px)

Citation preview

Page 1: The interrupted serosubmucosal anastomosis – still the gold standard

Original article

The interrupted serosubmucosal anastomosis – still the goldstandard

A. Leslie* and R. J. C. Steele*

*Department of Surgery and Molecular Oncology, University of Dundee, Dundee, UK

Received 8 July 2002; accepted 19 December 2002

Abstract

Background The single-layer appositional serosub-

mucosal anastomosis is a well established technique and

appears to have a favourable record. Over a 15-year

period the senior author of this paper has performed or

directly supervised 553 anastomoses using this technique.

This report describes the results of these operations, the

results of stapled anastomoses carried out during the

same period and discusses the utility of the handsewn

technique.

Methods From August 1986 to July 2001, 553 intes-

tinal anastomoses in 550 patients were fashioned using

single-layer, interrupted serosubmucosal 3 ⁄ 0 braided

polyamide and 131 anastomoses in 131 patients were

performed using a circular anastomosing stapler.

Results One anastomotic leakage occurred in the

group of patients whose anastomosis was handsewn

(0.2%) and 11 leaks occurred in those who had a stapled

anastomoses (8.4%). The mortality rate in each group was

similar (2% and 2.3%, respectively). There were no deaths

attributable to anastomotic dehiscence in either group.

Conclusion In this prospectively audited series of 553

handsewn anastomoses the leakage rate was 0.2%. These

results compare favourably with other published series

and continue to support a single layer of interrupted

serosubmucosal sutures as the gold standard for anasto-

moses involving the large or small bowel.

Keywords Serosubmucosal anastomosis, stapled anasto-

mosis, technique

Introduction

The concept of using the submucosal layer of the bowel

to hold stitches for a sound anastomosis was first

introduced by William Halsted in his publication of

1887 [1], but this described work in dogs, and there is no

evidence that Halsted ever applied his findings to man. It

fell to Matheson in Aberdeen, working from the 1960s to

the 1990s, to develop the technique for clinical use [2],

and the extramucosal, appositional, interrupted sero-

submucosal anastomosis has become widely used, with

reported leakage rates in the region of 2% [3,4].

Since 1986 the senior author of this paper has

employed this technique for all anastomoses involving

the large and small bowel with the exceptions of upper

gastrointestinal surgery, where a continuous sero-

submucosal technique was used [5], and low rectal

anastomoses, where stapling devices were employed. This

report describes the results of a continuous series of

operations carried out between August 1986 and July

2001 in four different centres either by, or under the

direct supervision of, the senior author with a view to

assessing the utility of the interrupted serosubmucosal

anastomosis. It represents an update of a previous report

[6], but on this occasion we have included all stapled low

anastomoses carried out by the same surgeon in order to

provide a complete picture of anastomosis-related mor-

bidity. It should be emphasized, however, that the

indications for sutured and stapled anastomoses in this

series were quite separate in that the stapled approach was

reserved for the lower anterior resections. This report

therefore should not be seen as a direct comparison of the

two techniques.

Patients and methods

Patients

Over a 15-year period 553 anastomoses involving the

large or small bowel in 550 patients (252 male, 298

Correspondence to: Ms A Leslie, Department of Surgery and Molecular Oncology,

University of Dundee, Ninewells Hospital, Dundee, DD1 9SY, UK.

E-mail: [email protected]

362 � 2003 Blackwell Publishing Ltd. Colorectal Disease, 5, 362–366

Page 2: The interrupted serosubmucosal anastomosis – still the gold standard

female) were fashioned using the interrupted sero-

submucosal technique. During the same period 131

anastomoses involving the colon and rectum in 131

patients (71 male, 60 female) were performed using a

circular anastomosing stapler. Both elective and emer-

gency cases are included in this series, and there are no

exclusions. The types of procedures, indications for

operation and complications are given in Tables 1 and

2. The term ‘Anterior Resection’ is used to describe

restorative excision of lesions at or below the recto-

sigmoid junction; an anastomosis at the pelvic brim is not

included in this category. Anastomotic leakage was

defined as any clinical evidence of leakage that led to

intervention, resulted in a prolonged hospital stay or was

diagnosed at postmortem. Post-operative death was

defined as death within 30 days of surgery.

A mid-line incision was used in all operations apart

from five laparoscopically assisted right hemicolectomies

and one laparoscopically assisted low anterior resection.

Peri-operative single-dose intravenous antibiotic prophy-

Table 1 Details of 553 single-layer interrupted serosubmucosal anastomoses in 550 patients.

Operation Indication No. of cases Complications and comments

Right hemicolectomy Cancer 109 1 death – MI. 1 anastomotic leak

Crohns’ disease 55 5 performed laparoscopically

Trauma 5

Ischaemia 2

Extended Right hemicolectomy Cancer 34 1 death – MI

Ischaemia 2

Colocutaneous fistula 1 1 death – whole gut ischaemia

Left hemicolectomy Cancer 31 2 deaths – MI & respiratory failure

Diverticular disease 1

Sigmoid colectomy Cancer 106 2 deaths – MI & PE

Diverticular disease 6 1 death – respiratory failure

Anterior Resection Cancer 48

Reversal of Hartmann’s Cancer 20 1 small bowel fistula

Diverticular disease 8

Trauma 3

Closure of loop colostomy – 11

Closure of loop ileostomy – 50

Subtotal colectomy & ileorectal anastomosis Cancer 10 2 deaths – adhesions & aspiration

FAP 5

Crohns’ disease 4

Constipation 3

Small bowel resection Crohns’ disease 16 1 death – MSOF

Trauma 6

Involved in colonic tumour 4

Ischaemia 3

Meckel’s diverticulitis 3

Gall stone ileus 3

Primary tumour 2

Involved in diverticular disease 2

MI, myocardial infarction; PE, pulmonary embolism; MSOF, Multi-system organ failure; TME, Total Mesorectal Excision.

Operation Indication No. of cases

Complications

and comments

High anterior resection Cancer 27 1 anastomotic leak

Low anterior resection &

TME with defunctioning

stoma

Cancer 104 3 deaths – 2 MI,

1 pneumonia

& respiratory failure.

10 anastomotic leaks

Table 2 Details of 131 stapled anasto-

moses in 131 patients.

A. Leslie & R. J. C. Steele Interrupted serosubmucosal anastomosis

� 2003 Blackwell Publishing Ltd. Colorectal Disease, 5, 362–366 363

Page 3: The interrupted serosubmucosal anastomosis – still the gold standard

laxis (cephalosporin and metronidazole) was used rou-

tinely and at the end of the procedure the peritoneal

cavity was lavaged with tetracycline solution or more

latterly, cephalosporin solution. The fascial incision was

closed with No.1 polydioxanone suture (PDS�: Ethicon,

Edinburgh, UK), and the skin was closed using a variety

of techniques.

Anastomotic technique

The technique for the appositional serosubmucosal

anastomosis is well described by Matheson in two

photographic colour atlases of right and left hemicolec-

tomy [7,8], and the principles described therein can be

applied to all but the most extreme situations. In all cases

in the present series a single layer of interrupted 3 ⁄ 0braided polyamide (NurulonR, Ethicon) was used for its

tying properties which allow controlled ‘snugging’ of the

knot on the second throw (Fig. 1). Care was taken to

include the submucosal and serosal layers in each suture,

but the mucosa was deliberately avoided (Fig. 2); sutures were placed an estimated 4 mm apart and 4 mm deep as

far as possible.

For mobile anastomoses, two corner sutures were

placed between the mesenteric and antimesenteric bor-

ders of both bowel ends and placed under tension by

means of the weight of Mayo’s artery forceps. A middle

suture was then placed exactly half way between the

corner sutures and again held with an artery forceps to aid

accurate placement of the remaining sutures which were

inserted towards the operator and held by the assistant.

On completion of the first side of the anastomosis the

sutures were tied by hand to ensure a snug but

nonconstricting knot, using, in the majority of cases, five

throws with the first two in the same direction to facilitate

controlled tightening. With the exception of the corners,

all the sutures were then cut to leave 4 mm long tails, the

bowel ends turned over, and the process repeated to

complete the anastomosis. It should be noted that the

corner sutures were not tied until the second line of

sutures had been inserted; this makes for easier identifi-

cation of the bowel wall layers at the corners.

For anastomoses that could not be turned over

(usually colorectal after sigmoid colectomy or high

anterior resection), the sequence of events was different

although the same basic principles of placement and tying

of the sutures were applied. Firstly, stay sutures were

placed through the serosubmucosal layers of the colon

equidistant between the mesenteric and antimesenteric

borders, and then through the rectum in the same

positions, all four being held by Mayo’s artery forceps.

The two corner sutures were then placed in the same

place and direction as the stay sutures, i.e. through the

serosubmucosal layer of the colon parallel to the cut edge

Figure 1 Tying the serosubmucosal suture, showing (a) the first

throw and (b) the second throw, both in the same direction to

form a slip knot. This is then locked with three further throws.Note the sutures placed approximately four millimetres apart and

four millimetres deep, and the degree of tension applied.

Figure 2 Insertion of the serosubmucosal suture, avoiding the

mucosa.

Interrupted serosubmucosal anastomosis A. Leslie & R. J. C. Steele

364 � 2003 Blackwell Publishing Ltd. Colorectal Disease, 5, 362–366

Page 4: The interrupted serosubmucosal anastomosis – still the gold standard

(not through it) about 4 mm from it, and then through

the rectal wall in the same way. This manoeuvre turns the

corners of the colon and rectum in when the sutures are

eventually tied, resulting in a neat result and making

placement of the adjacent sutures easier.

Once the corner sutures were inserted they were held

by Crile’s artery forceps which were hung over the edge

of the wound. A centre suture was then placed between

the midpoints of the posterior layers of the colon and

rectum and held by a further Crile’s forceps. Working

towards the operator, sutures were then inserted along

the posterior layers of the colon and rectum, each suture

being held in a forceps which was threaded in turn on to a

forceps holder to avoid tangling the threads. The corner

sutures were not held in the forceps holder. To facilitate

definition of the bowel layers, it was found to be helpful

for the assistant to hold up the previously inserted suture

under a slight degree of tension, and for the operator to

hold the centre or corner suture under similar tension

with a pair of tissue forceps. Once the posterior layer was

complete, the sutures in the forceps holder were held up,

the stay sutures removed, and the colon slid down to the

rectum. Each of the sutures was then tied, taking care to

tie the corner sutures square in order to turn the corners

in. After cutting all the sutures except the corners, the

anterior layer of sutures was then inserted and tied in a

similar fashion. For this anastomosis it has been found to

be useful to use the curved Heaney or Stratte needle

holder to obtain the correct angle of the needle.

It should be noted that in all cases meticulous care was

taken to ensure good blood supply to the cut ends of

bowel and complete lack of tension on the anastomosis.

Results

Following a total of 684 intestinal anastomoses, clinically

evident anastomotic leakage occurred in 12 patients

(1.8%). One leak occurred in the group of patients whose

anastomosis was fashioned using the interrupted sero-

submucosal technique (0.2%) and 11 occurred in the

group in whom a stapled anastomosis was performed

(8.4%). There were no specific features in these patients

that would have identified them at being at high risk of

anastomotic dehiscence.

There were a total of 14 deaths (2.1%): 11 in the

handsewn anastomosis group (2%) and three in the

stapled anastomosis group (2.3%). The cause of each

death is given in Tables 1 and 2. There were no deaths

associated with anastomotic dehiscence. In the handsewn

anastomotic group one patient developed a small bowel-

cutaneous fistula after reversal of a Hartmann’s operation;

this closed spontaneously. There were no other major

complications in either group.

Discussion

Published leakage rates after intestinal anastomosis are

highly variable. For example, in a recent meta-analysis of

13 randomised trials of handsewn vs stapled anastomosis

in colorectal surgery the clinical leakage rate for hand-

sewn varied from 0 to 10% and for stapled from 0 to 12%

[9]. Results from trials may not reflect experience

throughout the country, however, but owing to lack of

comprehensive prospective audit it is difficult to estimate

a reasonable average figure. Perhaps the most accurate

published data come from the Wales ⁄ Trent Audit, carried

out in 1992–93, which reported a 4% leakage rate after

colonic anastomosis and 8% after rectal anastomosis [10].

In this consecutive, prospectively audited series of 553

hand sewn anastomoses the leakage rate was 0.2%, a

figure which compares very well with any other large

published series, and, as far as we can tell, the national

average. There is little recent relevant literature other

than the meta-analysis referred to above, but a trial

comparing single-layer continuous with two layer inter-

rupted techniques published last year reported leakage

rates of 1.5% and 3.1%, respectively [11].

It has to be recognized that there is no universally

accepted and validated definition of anastomotic leak [12],

and this makes comparison of different series difficult.

However, the definition we adopted is in line with that

proposed by the UK Surgical Infection Study Group in

1991 [13], and we are confident that no clinically signi-

ficant event related to anastomotic leakage was missed.

Despite an impression that the interrupted sero-

submucosal anastomosis has become widely accepted

there are surprisingly few publications describing the

results of this technique. In Matheson’s own series,

published in 1985 [3], the leakage rate in 206 large

bowel intestinal anastomoses was 1.5%. Also in 1985

Errett et al. [14] described a 1.2% leak rate in 250 large

bowel anastomoses, and in the series reported by Carty

et al. [4] in 1991 it was 2.2% in 500 colorectal anastomoses.

Kingsnorth et al. [15] reported on a small heterogeneous

series in which they used both the open serosubmucosal

technique as described in this paper and a closed method.

They found the open method to be satisfactory with only

one leak out of 26 anastomoses, but the closed method was

associated with a leakage rate of 6.8% in 44 patients.

In our series of 553 handsewn anastomoses, one

clinically evident leak occurred in 1990. This was in a

27-year-old-man who had undergone a right hemicol-

ectomy for obstructing Crohn’s disease while on high-

dose steroids and cyclosporin. It is difficult to be sure

whether or not the therapy contributed to the leak, but

at re-laparotomy for peritonitis on day 2, the leakage

was traced to a single suture site. It was evident that

A. Leslie & R. J. C. Steele Interrupted serosubmucosal anastomosis

� 2003 Blackwell Publishing Ltd. Colorectal Disease, 5, 362–366 365

Page 5: The interrupted serosubmucosal anastomosis – still the gold standard

the suture had become untied, and from that time we

changed our policy to using five throws on every suture

as opposed to the three originally taught by Matheson.

The randomised trials comparing stapling with hand-

sewn anastomotic technique have not shown any differ-

ence in leakage rate [9], and the same is true of most trials

comparing one handsewn technique with another

[11,16–18]. It is likely therefore that the care with which

an anastomosis is carried out is more important than the

specific technique employed. There is no doubt, however,

that leakage rates can be kept very low by the interrupted

serosubmucosal technique, and while minimizing isch-

aemia by gentle apposition of the tissues may play a role,

we believe that the careful identification of the bowel

layers and the attention to detail demanded by the

technique ensures a sound result.

Although meticulous technique is clearly important,

anatomical location has a significant effect, and it is well

known that the low rectal anastomosis, particularly when

it is performed less that 5 cm from the anorectal junction

[19], is prone to leakage. This is borne out by our

experience of an 8.4% leak rate with the stapled anasto-

mosis after total mesorectal excision despite liberal use of

defunctioning stomas. Given the results of the random-

ised trials [9] these relatively poor results are not likely to

be due to stapling. This would be very difficult to prove

however; we have carried out moderately low rectal

anastomoses using the serosubmucosal technique, but

after total mesorectal excision we have found this to be

impossible, and the only way a manual anastomosis could

be achieved in this situation would involve the transanal

coloanal method. Although this technique has been used

after low anterior resection for cancer [20], it has never

been formally compared to a stapled anastomosis, and as

it would necessitate mucosectomy, it might have dele-

terious effects on function as experienced by some in

ileoanal pouch formation [21].

In conclusion, with the exception of low anterior

resection of rectum, leakage after intestinal anastomosis

can be virtually eliminated by the careful, standardized

use of a well-established technique. We commend the

interrupted serosubmucosal anastomosis as developed by

Matheson to all surgeons, particularly those who have an

interest in colorectal surgery.

References

1 Halsted WS. Circular suture of the intestine – an experi-

mental study. Am J Med Sci 1887; 94: 436–61.

2 Matheson NA. Prospective audit of an extramucosal tech-

nique for intestinal anastomosis. Br J Surg 1992; 79: 843.

3 Matheson NA, McIntosh CA, Krukowski ZH. Continu-

ing experience with single layer appositional anasto-

mosis in the large bowel. Br J Surg 1985; 72 (Suppl.):

S104–6.

4 Carty NJ, Keating J, Campbell J, Karanjia N, Heald RJ.

Prospective audit of an extramucosal technique for intestinal

anastomosis. Br J Surg 1991; 78: 1439–41.

5 Steele RJC. Continuous single-layer serosubmucosal anasto-

mosis in the upper gastrointestinal tract. Br J Surg 1993; 80:

1417.

6 Pye G, Steele RJC. Anastomoses involving the colon and

rectum: an 8-year experience. J R Coll Surg Edinb 1996; 41:

95–6.

7 Matheson NA. (1984) A Colour Atlas of Right Hemicolec-

tomy, pp. 47–63. Wolfe, London.

8 Matheson NA. (1986) A Colour Atlas of Left Hemicolectomy,

pp. 51–65. Wolfe, London.

9 McRae HM, McLeod RS. Handsewn vs. stapled anastomoses

incolonandrectal surgery.DisColonRectum1998;41:180–9.

10 Mella J, Biffin A, Radcliffe AG, Stamatakis JD, Steele RJC.

Population-based audit of colorectal cancer management in

two UK health regions. Br J Surg 1997; 84: 1731–6.

11 Burch JM, Franciose RJ, Moore EE, Biffl WL, Offner PJ.

Single-layer continuous versus two-layer interrupted intesti-

nal anastomosis: a prospective randomised trial. Ann Surg

2000; 231: 832–7.

12 Bruce J, Krukowski ZH, Al-Khairy G, Russell EM, Park

KGM. Systematic review of the definition and measurement

of anastomotic leak after gastrointestinal sugery. Br J Surg

2001; 88: 1157–68.

13 Peel AL, Taylor EW. Proposed definition for the audit of

post-operative infection: a discussion paper. Surgical Infec-

tion Study Group. Ann Roy Coll Surg Engl 1991; 73: 385–8.

14 Errett LE, Girotti M, Paloschi G. Single-layer large-bowel

anastomsis. a report of 250 cases. Can J Surg 1985; 28: 148–9.

15 Kingsnorth AN, Makin CA, Ellenbogen S. Prospective study

of the serosubmucosal (extramucosal) suture technique for

gastrointestinal anastomosis. J Roy Coll Surg Edin 1989; 34:

130–2.

16 Irvin TT, Goligher JC, Johnston D. A randomized pros-

pective clinical trial of single-layer and two-layer inverting

intestinal anastomoses. Br J Surg 1973; 60: 457–60.

17 Goligher JC, Lee PW, Simpkins KC, Lintott DJ. A

controlled comparison of one- and two-layer techniques of

suture for high and low colorectal anastomoses. Br J Surg

1977; 64: 609–14.

18 Ordorica-Flores RM, Bracho-Blanchett E, Nieto-Zermeno J

et al. Intestinal anastomoses in children: a comparative study

between two different techniques. J Paediatric Surg 1998;

33: 1757–9.

19 Rullier E, Laurent C, Garrelon JL, Michel P, Saric J, Parneix

M. Risk factors for anastomotic leakage after resection for

rectal cancer. Br J Surg 1998; 85: 355–8.

20 Enker WE, Sterns MW, Janov AJ. Peranal coloanal anasto-

mosis following low anterior resection for rectal carcinoma.

Dis Colon Rectum 1985; 28: 576–81.

21 Gemlo BT, Belmomte C, Wiltz O, Madoff RD. Functional

assessment of ileal pouch-anal anastomotic techniques. Am J

Surg 1995; 169: 137–41.

Interrupted serosubmucosal anastomosis A. Leslie & R. J. C. Steele

366 � 2003 Blackwell Publishing Ltd. Colorectal Disease, 5, 362–366