Ex vivo comparison of three hand sewn end-to-end anastomoses in normal equine jejunum


  • Ceri Sherlock's present address is: The School of Veterinary Medicine and Science, The University of Nottingham, Sutton Bonington Campus, Leicestershire LE12 5RD, UK.

email: cerisherlock@hotmail.com


Reasons for performing study: Although single layer techniques are preferred in man and small animals for small intestinal end-to-end anastomoses, double layer techniques are more popular in equine surgery. This study was undertaken to evaluate the ex vivo characteristics of 2 single layer anastomoses in comparison to the traditionally performed double layer anastomosis in equine jejunum.

Objectives: To compare ex vivo: 1) construction time; 2) bursting pressure; and 3) lumen size reduction of 3 suture patterns (double layer simple continuous/Cushing pattern [2C]; single layer Lembert pattern [1L]; and single layer Cushing pattern [1C]) for end-to-end equine jejunojejunostomies.

Methods: End-to-end jejunojejunostomies were constructed using 2C (n = 7), 1L (n = 7) and 1C (n = 7) in harvested equine jejunum and construction times were recorded. Anastomosed and control segments were distended with gas until failure. Intraluminal pressure at failure and mode of failure were recorded. Lumen size reduction was calculated as a percentage decrease from control jejunum. Results were compared using an ANOVA and P<0.05 was considered significant.

Results: The 1C anastomoses were faster to construct than the 1L anastomoses, which were faster to construct than the 2C anastomoses. There were no differences in bursting pressures between the different anastomoses and control jejunum. All anastomoses decreased lumen size from control values but there were no differences in lumen reduction between types of anastomoses.

Conclusions: Single layer anastomoses are faster to construct than double layer anastomoses, with the 1C being fastest. Single layer anastomoses are as strong and result in comparable lumen size reduction as traditional 2C anastomoses.

Potential relevance: As the 1C anastomosis results in less exposed potentially adhesiogenic suture material than the 1L while providing adequate strength and similar luminal size reduction, the 1C may be better for equine small intestine anastomosis and further in vivo studies are warranted.


Lesions requiring resection and anastomosis account for 58–80% of disorders of the small intestine in horses undergoing exploratory celiotomy (Freeman 1997; White 1999; Latson et al. 2005). Many surgical techniques have been reported for jejunojejunal anastomosis including end-to-end, functional end-to-end and side-to-side methods. Although no difference in complication rates between techniques were detected in some reports of clinical cases (Semevolos et al. 2002) or experimentally (Baxter et al. 1992), the end-to-end technique was found to have a lower risk of complications in comparison to the functional end-to-end anastomosis technique (Franheny et al. 1995) and a lower cost in comparison to the stapled side-to-side technique (Freeman 2003). Additionally, the end-to-end technique is the most physiological way for the intestine to be apposed. For these reasons, many surgeons prefer the end-to-end technique.

The end-to-end technique is generally performed by hand suturing the resected ends of the intestine together as end-to-end stapling devices are too small for use in mature equine small intestinal anastomosis (Freeman 1997). Many methods of hand suturing the intestine have been described and these include interrupted and continuous patterns, inverting, appositional and everting patterns, and patterns in a double or single layer (Reinertson 1976; Freeman 1997, 2003; Eggleston et al. 2004). Interrupted patterns avoid a purse string effect and may allow some expansion during passage of a food bolus; however, they are slow to perform and expose large amounts of suture that can predispose to adhesion formation (Freeman 2003). Everting suture patterns predispose to adhesion formation and peritonitis (Reinertson 1976). Appositional patterns are also reported to predispose to adhesion formation (Eggleston et al. 2004). Currently, the most common methods for end-to-end anastomoses are single or double layer closures with the most superficial layer an inverting pattern. Although single layer techniques are preferred in human and small animal intestinal anastomoses (Coolman et al. 2000), double layer techniques are more popular in equine surgery (Freeman 1997) due to the perceived reduction in anastomotic complications including adhesion formation and anastomotic leakage.

A recent ex vivo study demonstrated that a single layer Lembert pattern (1L) was quicker to construct, created a larger lumendiameter but equivalent bursting strength for jejunojejunal end-to-end anastomosis compared with a double layer pattern (full thickness simple continuous and subsequent Cushing pattern) (Nieto et al. 2006). A single layer Cushing pattern (1C) may be superior to 1L because of decreased exposed potentially adhesiogenic suture material, a short construction time, equivalent bursting pressure and maximal lumen size.

The objectives of this study were to compare properties of 3 suture patterns; traditional double layer, simple continuous mucosa-submucosa oversewn with Cushing pattern (2C), 1L and 1C for end-to-end equine jejunojejunal anastomoses ex vivo. Properties that were compared were time of construct, maximum bursting pressure, bursting wall tension and lumen size reduction. We hypothesised that the 1C would be quickest to construct, maintain comparable bursting pressure and bursting wall tension, and have less lumen size reduction than the other suture patterns ex vivo.

Materials and methods

Twenty-eight 30 cm mid jejunal segments were harvested from 7 mature horses (age range 3–28 years; 3 geldings and 4 mares) immediately after euthanasia. All horses were subjected to euthanasia for reasons unrelated to diseases of the gastrointestinal system. Intestinal segments were evacuated of luminal contents, lavaged and then immersed in 0.9% saline solution at room temperature. Anastomosis construction and testing were performed within 4 h of euthanasia.

Anastomosis technique

One of each of the 3 types of anastomoses (2C, 1L, 1C) was performed in each horse by a single surgeon, who is a diplomate of the ACVS (C.S.), and one segment from each horse served as a control. For all anastomoses, the small intestinal segments were placed flat and sharply transected at 60° to the mesenteric attachment. Stay sutures were placed in the mesenteric and anti-mesenteric borders to appose the transected segments. Segments were mounted on a convex laminated saline lavaged surface and 30 g weights (Allis tissue forceps) were attached to stay sutures to supply uniform tension at every anastomosis site. Segments were lavaged with 0.9% saline solution intermittently during anastomosis construction. Total construction time was recorded as the time from completion of stay suture placement and tension application to completion of the anastomosis. Polydioxanone (3-0 PDS II)1 and a swaged 26 mm half circle taper point needle were used for all anastomoses. All patterns and layers were interrupted at the mesenteric and antimesenteric borders to prevent a purse string effect. Adequate suture line tension was applied to induce inversion of the serosal layer. Any areas of knot exposure or questionable serosal inversion were oversewn in an interrupted inverting pattern as performed by the authors in clinical cases.

Double layer simple continuous and Cushing pattern (2C)

The mucosal-submucosal layer was sutured in a simple continuous pattern. Suture bites were taken 3–5 mm apart and 1–3 mm from the incision edge. This was then oversewn with a continuous Cushing pattern in the seromuscular and submucosal layers with suture bites 2–3 mm from the incision and approximately 5 mm long.

Single layer Lembert pattern (1L)

The seromuscular and submucosal layers were sutured in a continuous Lembert pattern. Suture was placed 1 mm from the incised edge and bites were approximately 5 mm long.

Single layer Cushing pattern (1C)

The seromuscular and submucosal layers were sutured in a continuous Cushing pattern and suture bites were placed as for the Cushing component of the 2C.

Biomechanical testing

Bursting pressure (BP, mmHg) and bursting wall tension (BWT, dynes/cm) of intestinal segments were determined using an apparatus described by Eggleston et al. (2004) for anastomosed and control segments from each horse. Briefly, the intestinal segments were sealed at both ends by an extraluminal ligature, the lengths were recorded and the segments were submerged in room temperature 0.9% saline solution within a chamber. Room air was infused into the intestinal segments and intraluminal pressure was recorded continuously on a chart recorder (LabChart7)2 from a calibrated pressure transducer and data acquisition system (PowerLab 8/30)2. Maximum pressure at which failure occurred (BP) and mode of failure were recorded. The volume of fluid displaced from the chamber was also recorded. The BWT was subsequently calculated using Laplace's law (Bickers et al. 2002; Eggleston et al. 2004; Nieto et al. 2006) using the following formula: BWT = BP × r, where BP is in dynes/cm2 (1 mmHg = 1333.22 dynes/cm2) and r is the radius of the intestinal segment (cm). The radius was calculated presuming that the segment was a cylinder by using the following equation: r =√(v/πl) where v is the volume (ml) of fluid displaced during intestinal distension until failure and l is the recorded length of the segment(cm).

Luminal diameter reduction

The same intestinal segments that underwent biomechanical testing were transected parallel and immediately adjacent to one side of the anastomosis. After mounting the anastomosis on the convex laminated saline lavaged surface and applying uniform tension via weights on stay sutures as previously described for anastomosis construction, Vernier calipers were used to measure the length of the anastomosis (the distance from the mesenteric border to the anti-mesenteric border of the stretched bowel) to represent anastomosis size. Control length of lumen was measured by transecting an intestinal segment half way between the intestinal seal and the anastomosis for each anastomosed segment and mounting and measuring as for the anastomosis lumen measurement above. The final control measurement was an average of the control measurements for each horse. Luminal reduction at the anastomosis was expressed as a percentage of the final control measurement in order to decrease the effect of interhorse variability in jejunal lumen diameter.

Statistical analysis

Data were tested for normality using a Shapiro-Wilk test. Results were reported as mean ± s.e. A one-way repeated measures ANOVA with a post hoc pairwise multiple comparison procedure (Holm-Sidak method) was used to compare construct time, bursting pressures, bursting wall tension and luminal reduction between anastomotic techniques. Statistical significance was established at P<0.05.


Anastomosis construction time

Construction time was significantly different between groups (P≤0.001) (Table 1). The 1C was significantly faster than the 1L and the 2C anastomoses. The 1L was significantly faster than the 2C anastomoses.

Table 1. Mean ± s.e. values of anastomosis construction time, bursting pressure (BP) (mmHg), bursting wall tension (BWT), and lumen size reduction (%) in mid jejunal segments from 7 horses after end-to-end anastomosis in a single layer Cushing pattern (1C), a single layer Lembert pattern (1L) or a double layer mucosa-submucosa simple continuous followed by a Cushing pattern (2C)
GroupConstruction time (min)BP (mmHg)BWT (dynes/cm)Lumen reduction (%)
  • a,b,c

    Values marked with superscript letters are significantly different from each other (P<0.05).

Controln/a290.60 ± 35.061,698,448 ± 212,558n/a
1C17.16 ± 0.48a251.56 ± 56.791,141,829 ± 197,12417.00 ± 2.19
1L20.34 ± 0.94b222.61 ± 38.231,114,829 ± 233,47815.33 ± 2.06
2C23.67 ± 0.76c228.77 ± 42.761,087,833 ± 319,05220.29 ± 2.71

Biomechanical testing

There were no differences in BP or BWT between anastomotic techniques or control segments (Table 1). Segment failure in the control group occurred at the mesentery in 6 segments and at the intestinal seal in one segment (the tissue slid from the ligature). Segment failure in the anastomoses groups occurred at the mesentery away from the anastomosis in 6 segments (1C, n = 2; 1L, n = 2; 2C, n = 2). The remainder of the anastomosed segments failed by leakage through the anastomosis at the mesenteric (1C, n = 3; 1L, n = 3; 2C, n = 3) or antimesenteric borders (1C, n = 2; 1L, n = 2; 2C, n = 2). All anastomosed segments demonstrated serosal tearing at the point of at least one suture penetration.

Luminal size reduction

Lumen size was reduced compared to control values with all anastomotic techniques (Table 1). There was no significant difference in the magnitude of luminal reduction between different anastomotic techniques.


Results from this ex vivo study indicate that single layer anastomoses (1C and 1L) are significantly quicker to perform than double layer anastomoses (2C). Furthermore, the 1C pattern is significantly quicker to perform than the 1L pattern. All 3 suture patterns provide comparable BP, BWT and lumen size reduction.

The decreased construction time for single layer anastomoses has been noted previously with double layer anastomoses having a mean ± s.e. of 23.68 ± 0.75 min similar to times in this study, and the 1L taking 17.55 ± 0.96 min which is slightly quicker than the times for 1L in this study (Nieto et al. 2006). The results of this study therefore confirm the decreased construction time for single layer techniques compared with the traditionally performed double layer techniques, which is likely to be an advantage in vivo. Although it is acknowledged that the mean construction time difference between 1C and 2C was only 6 min, any factor contributing to shortening surgery time is worthy of consideration in clinical cases as there is a negative correlation between surgery time and survival in horses undergoing exploratory celiotomy (Phillips and Walmsley 1993; Proudman et al. 2005). Decreased anastomosis construction time will also reduce manipulation of the intestine and therefore the resultant iatrogenic inflammation (Little et al. 2005) that predisposes to adhesion formation (Baxter 1991), and may contribute to post operative ileus (Little et al. 2005).

No difference in BP or BWT was found between the different anastomotic techniques. Additionally, the BP and BWT for segments with an anastomosis were similar to normal intestine. The intraluminal pressures reached in this study far exceed those reported to cause ischaemic damage to the equine mucosa experimentally (18 cmH2O [13.2 mmHg] for 120 min) (Dabareiner et al. 2001), those noted in nonsurviving clinical cases with small intestinal obstruction (average 15 cmH2O [11 mmHg]) (Allen et al. 1986), and those noted in human subjects with experimentally obstructed small intestine (50 cmH2O [36.78 mmHg]) (Abbott et al. 1943). Therefore, anastomoses constructed using these suture patterns are unlikely to fail in clinical cases due to excessive intraluminal pressures.

The most common method of anastomosis segment failure in this study was different from previously reported methods of failure (Nieto et al. 2006). In this study, most anastomosed segments failed by leakage of small bubbles through the anastomosis. However, the BP appeared similar to those reported previously (Bickers et al. 2002; Eggleston et al. 2004; Gandini 2006; Gandini and Bertuglia 2006; Nieto et al. 2006). We attribute the differences in recognised failure methods to the easy identification of small bubbles of air within fluid, in comparison to the more subtle identification of leakage of small volumes of coloured fluid into a water bath (Nieto et al. 2006); however, other methodological differences may have contributed to these disparities including the medium and the flow rate used to produce intestinal distension as BP measurements can vary with the speed of distension (Nelsen and Anders 1966). All suture lines demonstrated the presence of some serosal tearing irrespective of the ultimate method of segment failure in this study. This has been noted similarly in small intestinal suture lines in other species (Semevolos et al. 2007). The serosal tearing in combination with an intact suture line emphasises the importance of inclusion of the holding layer (the submucosa) in the suture lines of anastomoses (Halsted 1887). The submucosa reportedly contributes 70–75% of the strength of human intestinal wall while the serosa is only reported to contribute to 5–10% of the strength (Raikevitch 1963). The submucosa provides the mechanical stability, strength and resistance to intensive deformations of long duration in human small intestine (Egorov et al. 2002). Inclusion of the submucosa within each suture bite whilst avoiding suture bites of excessive length predisposing to overinversion may be more challenging in live horses whose intestine is likely to swell in response to handling; the surgeon must therefore use their discretion to ensure that bites attain sufficient depth without excessive inversion.

Creation of a ‘purse-string’ effect and subsequent anastomotic stenosis with 1C (Freeman 2003) was avoided in this study; there was no significant difference in lumen size reduction between the 3 suture patterns. The absence of a ‘purse-string’ effect may have been due to careful attention to its prevention that should be adhered to in vivo. In contrast to our results, a previous study reported a larger anastomosis lumen with a 1L in comparison to a double layer simple continuous and Cushing pattern (Nieto et al. 2006). This difference may be attributed to the simple continuous layer being full thickness in the Nieto et al. (2006) study, while the simple continuous layer was partial thickness (mucosa and submucosa only) in this study. The method of construction of 2C anastomosis was chosen because it is the technique used by the authors clinically. It was surprising that there was no significant difference in the lumen reduction between the 1C and 1L patterns as 1L creates more inversion (Freeman 2003); this may be attributed to the careful technique used to prevent over-inversion with the Lembert pattern as is recommended in vivo. In future studies, alternative methods of lumen size measurements are worthy of consideration; measurements taken during fluid distension to a uniform pressure may have produced more consistent results between samples (Eggleston et al. 2001; Semevolos et al. 2007). Additionally, the power of this statistical comparison was low and measurement of additional anastomoses may be required to ensure that no significant difference exists.

Creation of the largest lumen diameter is a goal of small intestinal anastomosis construction to prevent stricture formation. In this study, the lumen size may not have been maximised because the authors noted that the tension placed on the stay sutures was subjectively less than that placed on stay sutures in clinical cases; however, more tension was placed on these stay sutures than has been reported previously in similar studies (Nieto et al. 2006). Reduction in lumen diameter may progress as the anastomosis heals. An average of 50% reduction in luminal diameter 30 days after construction of a 2C anastomosis was found previously (Baxter et al. 1992). In order to determine the effect of healing on lumen diameter reduction, in vivo evaluation of the 1C and 1L anastomoses would be required.

In conclusion, the findings of this ex vivo study suggest that single layer techniques for anastomosis of the equine small intestine are biomechanically equivalent to the traditionally performed double layer anastomosis. This is supported by a recent clinical report of 15 horses undergoing small intestinal 1L anastomoses that demonstrated comparable short- and long-term survival rates (Mendez-Angulo et al. 2010) to horses undergoing 2 layer anastomoses reported previously (Semevolos et al. 2002; Proudman et al. 2007). This study has also demonstrated that the single layer 1C anastomosis is faster to construct than the 1L and 2C anastomoses. The 1C anastomosis has potential advantages over the 1L anastomosis as it is technically easy and results in less exposed suture material that may predispose to adhesion formation (Freeman et al. 2000). The advantages of faster construction and thereby decreased bowel manipulation during construction, and decreased exposed suture material with the 1C jejunojejunostomy warrant further in vivo studies.

Authors' declaration of interests

The authors have no conflicts of interest.

Source of funding

This study was funded the ‘For the Love of the Horse Equine Endowment Fund’, The University of Georgia.


We express our gratitude to Tom Robertson and Tristan Lewis who assisted with biomechanical testing.

Manufacturers' addresses

1 Ethicon Inc., Somerville, New Jersey, USA.

2 ADInstruments, Inc., Colorado Springs, Colorado, USA.