Difference in incisional complications following exploratory celiotomies using antibacterial-coated suture material for subcutaneous closure: Prospective randomised study in 100 horses

Authors


Summary

Reasons for performing study: Incisional complications are a major post operative challenge following ventral midline exploratory celiotomy for abdominal pain in horses. They lead to discomfort, prolonged hospitalisation, longer recovery times and increased cost; therefore, investigation of preventative procedures are warranted.

Objectives: To determine the clinical effect of antibacterial (triclosan) coated 2-0 polyglactin 910 suture material on the likelihood of incisional infections when used for closure of subcutaneous tissue following ventral midline celiotomies in horses.

Methods: One hundred horses undergoing exploratory celiotomy assigned at random to one of 2 groups. In the control group coated 2-0 polyglactin 910 (Vicryl) was used for apposition of the subcutaneous tissue in a simple continuous pattern and, in the study group, antibacterial (triclosan) coated 2-0 polyglactin 910 suture material (Vicryl Plus) was used. Post operatively an elastic adhesive abdominal bandage was applied, changed and the incision was inspected by a clinician blinded to the study protocol at 24–36 h and 6–9 days post operatively. Outcomes of interest were evidence of incisional pain, incisional oedema, sheath/udder oedema, incisional drainage, hernia formation and dehiscence.

Results: Antibacterial-coated suture material did not decrease the likelihood of incisional complications in 100 horses.

Conclusions: A beneficial effect on ventral midline incisions in 100 horses was not evident by using antibacterial-coated suture material.

Potential relevance: Lack of effectiveness of antibacterial-coated suture material in equine ventral midline closure after exploratory celiotomy and the observed potential adverse effects suggest that further clinical investigations are needed before using such material routinely on horses.

Introduction

Incisional complications, including oedema, drainage, infection, herniation and dehiscence, belong to the most common complications following ventral midline celiotomy in cases of abdominal pain in horses and lead to discomfort, prolonged hospitalisation, longer recovery times and increased cost. Incisional complication rates have varied widely in previous studies: 37% (Phillips and Walmsley 1993), 28.6% (Kobluk et al. 1989), 26.9% (Mair and Smith 2005a), 25.4% (Honnas and Cohen 1997), 16% (Proudman et al. 2002) and 7.4% (Freeman et al. 2000). Incisional drainage is considered to be indicative of incisional infection and was shown to be significantly associated with the occurrence of more serious incisional complications, such as hernia formation (Gibson et al. 1989; French et al. 2002; Mair and Smith 2005b).

Many studies have identified surgical and case risk factors that contribute to the likelihood of incisional complications. These have included heart rate and endotoxaemia at presentation (French et al. 2002), use of polyglactin 910 suture material (Honnas and Cohen 1997) or a near-far-far-near suture pattern for closure of the linea alba (Kobluk et al. 1989), 2-layer vs. 3-layer closure of ventral midline incisions (Coomer et al. 2007), performing a routine enterotomy or intestinal resection (Phillips and Walmsley 1993; Honnas and Cohen 1997), caecum or large colon obstruction (Phillips and Walmsley 1993), duration of surgery (Wilson et al. 1995), incisional contamination during anaesthetic recovery (Galuppo et al. 1999) and increased concentration of fibrinogen in peritoneal fluid prior to surgery (Honnas and Cohen 1997). Mair and Smith (2005a) found that multiple factors significantly influence the occurrence of incisional complications. The rate of complications was increased when total plasma protein concentration was increased and the linea alba was dissected prior to closure. Factors such as administration of intraperitoneal heparin or application of wound coverage decreased the likelihood of incisional complications.

Only a few previous studies have evaluated specific strategies to reduce the incidence of incisional complications. There was anabsolute risk reduction of the likelihood of developing a post operative incisional complication of 45% when using an abdominal bandage compared to not using one in the post operative period (Smith et al. 2007). Results of data at the Chino Valley Equine Hospital, showed that the use of a hernia belt post operatively significantly decreased the incidence of incisional hernia formation after midline exploratory celiotomy (Klohnen et al. 2007).

The role of surgical sutures in the aetiology of surgical site infection has been investigated (Edmiston et al. 2006). The prevention and treatment of surgical site infections is one of the main surgical challenges. In vitro and in vivo studies have shown that antibacterial-coated suture material may play a beneficial role in the prevention of post surgical infection. In vitro zone of inhibition assays showed that antibacterial (triclosan) coated 2-0 polyglactin 910 suture material (Vicryl Plus)1 provides an antimicrobial effect sufficient to prevent colonisation by Staphylococcus aureus and Staphylococcus epidermis (Rothenburger et al. 2002). In vivo studies in animal models demonstrate that antibacterial-coated suture material inhibits bacterial colonisation of suture after direct in vivo inoculation challenge with Staphylococcus aureus in a guinea pig model (Storch et al. 2004). In rats, the deep zone of a surgical site was contaminated with Staphylococcus epidermis and the incision closed with antibacterial-coated suture material. When using antibacterial suture material the number of positive cultures after surgery was reduced by 66.6% (Marco et al. 2007). In vivo studies in paediatric patients suggest that the incidence of post operative pain was significantly less and there was a trend toward lower incidence of incisional oedema when treated with antibacterial-coated suture material (Ford et al. 2005). The clinical efficacy of antibacterial-coated suture material has been shown by Fleck et al. in human cardiothoracic patients. While 24 of 376 patients closed with a nonantibacterial suture developed complications, none of the 103 wounds closed with triclosan-coated suture material developed infection or dehiscence during hospital stay and follow-up visits (Fleck et al. 2007). Results suggest that antibacterial-coated suture material constitutes a new weapon in the fight against post operative infection, particularly in hernioplasty, emergency surgery and dirty or potentially contaminated surgery. To the authors' knowledge there is no controlled clinical study reported in veterinary medicine and therefore the clinical significance of these findings in veterinary medicine, particularly in equine cases, remains questionable.

The objective of the present study was to perform a prospective, randomised and blinded study to determine the clinical effect of antibacterial-coated 2-0 Polyglactin 910 (Vicryl Plus) on the likelihood of incisional complications compared to 2-0 polyglactin 910 suture material (Vicryl)1 following ventral midline exploratory celiotomies in 100 horses with abdominal pain. It was hypothesised that the antibacterial effect of triclosan may decrease the likelihood of incisional complications following ventral midline exploratory celiotomy in horses and have no adverse effects.

Materials and methods

One hundred horses undergoing exploratory celiotomy through the linea alba at the Chino Valley Equine Hospital (Pomona, California, USA) were included in the study. Criteria for inclusion in the study were successful recovery from anaesthesia and not subjected to euthanasia prior to staple removal 10 days post operatively. Data were recorded at 4 time points: preoperatively, intraoperatively, 24–36 h post operatively and 6–9 days post operatively.

Preoperative variables including age, sex, weight of horse; heart rate, respiratory rate, temperature, mucous membrane colour and capillary refill time at presentation as well as duration of colic, degree of abdominal pain at presentation on a scale of 1–3 (1 = pawing; 2 = rolling; 3 = vigorously rolling, nonresponsiveness to analgesia) and basic haematological and biochemical parameters (packed cell volume, total protein, white blood cell count, fibrinogen, creatinine, electrolytes (K+, Na+, HCO3-, Ca2+) were recorded. Preoperatively, horses were medicated with broad-spectrum antimicrobial drugs (procaine penicillin 22,000 iu/kg bwt i.m. and gentamicin 6.6 mg/kg bwt i.v.) and nonsteroidal anti-inflammatory drugs (flunixin meglumine 1.1 mg/kg bwt i.v.). Post operatively, they were subsequently treated for 3–5 days (procaine penicillin 22,000 iu/kg bwt i.m. b.i.d. and gentamicin 6.6 mg/kg bwt i.v. s.i.d.) and 10 days (flunixin meglumine 1.1 mg/kg bwt i.v. b.i.d.), respectively. The surgeon determined the duration of antibiotic treatment depending on the surgical procedure performed. Horses with resections and enterotomies received 5 days of antimicrobial drugs. Horses without enterotomies were treated with 3 days of antimicrobial drugs.

Intraoperative data regarding duration of general anaesthesia, surgeon, anatomical location and primary lesion identified, performance of an enterotomy or intestinal resection with anastomosis and surgical incision length were recorded. The ventral abdomen was shaved immediately preoperatively and the skin was prepared aseptically with povidone-iodine, water and alcohol. The skin was dried with a sterile towel and sterile gauze. The abdomen was subsequently covered with an iodophor impregnated adhesive drape and covered with a water-impervious drape that included an adhesive lined fenestration positioned over the plastic adhesive drape.

All horses received 1 g of cephazolin2, diluted in one litre of sterile saline, intra-abdominally prior to closure. The linea alba was closed in an interrupted inverse cruciate pattern using 3 polyglactin 910 suture material and lavaged with cephazolin solution and saline prior to apposition of subcutaneous and skin layers using antibacterial-coated or noncoated 2-0 polyglactin 910 suture material. Intraoperatively the horse was assigned randomly to the study group (antibacterial-coated suture material) or the control group (nonantibacterial-coated suture material) to determine which suture material was used for the apposition of the subcutaneous tissue in a simple continuous pattern. Randomisation was achieved by tossing a coin intraoperatively. The skin was apposed using stainless steel staples. For the recovery period, an iodophore impregnated self-adhesive drape was applied to protect the incision. After successful recovery, the iodophore drape was removed and an abdominal bandage consisting of elastic adhesive bandages and 2 Gamgee cotton pads (one placed on the ventral midline and one on the dorsal spine to prevent pressure sores) were applied.

The abdominal bandage was changed 24–36 h post operatively and the incision inspected. If a horse became pyrexic in the post operative period, the bandage was removed and the incision was reinspected. Routinely the bandage was removed 6–9 days post operatively and the incision reinspected. If there was no post operative incisional complication the staples were removed 10 days after surgery and the horse discharged from the hospital, as allowed, dependent upon their medical condition. Evidence of incisional oedema or sheath/udder oedema (none/slight/moderate/severe), incisional drainage (none/slight/moderate/severe), pain (none/slight/moderate/severe), dehiscence (none/skin/subcutaneous tissue/body wall) and hernia formation (yes/no) were noted. Hospitalisation time (days) and outcome (discharged/euthanasia) were recorded. The clinician evaluating the incision post operatively was unaware which suture material had been used intraoperatively.

Statistical analysis

Descriptive statistics for all cases and variables were performed and recorded in percentages or mean ± s.d. Statistical analysis was performed with SPSS 6.0 computer software3. Quantitative variables that could bias the outcome (age, weight, duration of colic, temperature, heart rate, respiration rate, incision length, anaesthesia time and bloodwork abnormalities) of the study and control group and adverse events at first and second bandage were tested with independent t tests and analysis of variance. Difference of proportion tests was used to analyse adverse events (incisional pain, incisional oedema, sheath/udder oedema, amount of drainage, dehiscence and herniation) at first and second bandage change comparing study and control group. The associations between categorical variables (degree of abdominal pain at presentation, surgeon, performance of intestinal resection or enterotomy and anatomical location and primary lesion identified) and adverse events at first and second bandage change were analysed using Mantel–Haenszel test of linear association and the Goodman and Kruskal's tau statistic. Significance level was set at P<0.05.

Results

One hundred horses met the inclusion criteria: 47 were assigned to the study group and 53 to the control group. Descriptive statistics for all cases were recorded (Table 1). Cases in which large colon displacements, enteroliths and faecaliths were identified as primary lesions generally had secondary large colon impactions. Primary large colon impactions were only identified as such if there were no other bowel disorders present at exploration of the abdomen. One horse had 2 enteroliths in the large colon and one in the small colon and was counted in the large colon group. The mean ± age of the study population was 10.6 ± 7.5 years and mean bwt 498 ±  72.7 kg; mean duration of colic prior to presentation was 19 ±  40.9 h. One horse that had a nephrosplenic entrapment of the large colon had duration of colic of 16.5 days. Mean anaesthesia time was 135 min ± 40.5 and mean incisional length 27.5 ±  4.5 cm. The mean incision length for horses experiencing no oedema at the first bandage change was significantly shorter (mean = 27 cm, n = 92) than for horses experiencing moderate incisional oedema (mean = 31.5 cm, n = 7) (P = 0.033). Ninety-two cases were discharged from the hospital and 8 were subjected to euthanasia after mean hospitalisation time 18 days ± 15.7. The 8 horses subjected to euthanasia had their staples removed 10 days post operatively and were included in the results.

Table 1. Descriptive statistics for cases treated with antibacterial-coated suture material (study group) and non-antibacterial-coated suture material (control group)
 Study group
no./%
Control group
no./%
Gender  
 Male44
 Male castrate2631
 Female1718
Breed  
 Warmblood98
 Quarter Horse812
 Thoroughhbred1317
 Arabian Horse85
 Andalusian/Friesian33
 Draft Horse11
 Other57
Level of abdominal discomfort at presentation  
 Grade 1 (Pawing)1610
 Grade 2 (Rolling)2220
 Grade 3 (Rolling, nonresponsive to analgesia)923
Surgeon  
 A2829
 B1924
Anatomic location of primary lesion  
 Large intestine3736
 Small intestine1017
Primary lesion identified  
 Large colon torsion137
 Large colon displacement711
 Nephrosplenic entrapment of large colon24
 Large colon impaction (sand, gravel, ingesta)51
 Large colon faecalith13
 Large colon enteroliths34
 Small colon impaction (ingesta)12
 Small colon faecalith12
 Small colon enterolith30
 Small colon foreign body01
 Strangulating lesion of small colon10
 Strangulating lesion of small intestine1011
 Nonstrangulating lesion of small intestine11
 Entrapment of small intestine in epiploic foramen02
 Inflammatory infiltrate of bowl wall03
Resection/enterotomy performed  
 Large colon20
 Small intestine613
 Small colon10
 Full thickness biopsy01
 Pelvic flexure enterotomy4042

Age (P = 0.839), weight (P = 0.623), duration of colic (P = 0.321), temperature (P = 0.623), heart rate (P = 0.420), respiratory rate (P = 0.571), anaesthesia time (P = 0.395) and preoperative bloodwork values did not differ significantly between study and control groups.

Incisional complications at first and second bandage change were recorded for all cases (Table 2). Overall, 26% of all cases had incisional drainage. At the first bandage change one horse in the study group had incisional drainage. At the second bandage change 12 horses in the control group and 14 horses in the study group showed drainage. At the first bandage change (24–36 h post operatively) antibacterial-coated suture material had no significant effect on incisional pain (P = 0.289), incisional oedema (P = 0.099), sheath/udder oedema (P = 0.289), incisional drainage (P = 0.289), hernia formation (P = 0.289) and dehiscence (P = 0.289). At the second bandage change (6–9 days post operatively) antibacterial-coated suture material did not show any significant effect on adverse events: incisional pain (P = 0.93), incisional oedema (P = 0.081), sheath/udder oedema (P = 0.138), incisional drainage (P = 0.92), hernia formation (P = 0.289) and dehiscence (P = 0.87).

Table 2. Incisional complications recorded at first and second bandage change for cases treated with antibacterial-coated suture material (study group) and non-antibacterial-coated suture material (control group)
 1st bandage change2nd bandage change
Study group
no./47
Control group
no./53
Study group
no./47
Control group
no./53
Pain    
 No47533641
 Yes001112
Incisional oedema    
 No415186
 Slight522438
 Moderate10139
 Severe0020
Sheath/udder oedema    
 No46533041
 Slight0089
 Moderate1063
 Severe0030
Amount of drainage    
 No46533539
 Mild0077
 Moderate0044
 Severe1013
Quality of drainage    
 No46533539
 Purulent10811
 Serosanguineous0043
Dehiscence    
 No46534348
 Skin0010
 Subcutaneous tissue0025
 Body wall1010
Hernia    
 No46534653
 Yes1010

Variables, such as surgeon, degree of abdominal pain at presentation, performance of intestinal resection, performance of an enterotomy, anatomical location and primary lesion identified as predictors of adverse events, did not show any significant associations.

Discussion

This study shows that there was no evidence of effectiveness of antibacterial-coated suture material to prevent incisional complications in ventral midline celiotomy closure in 100 horses.

The antimicrobial effect of triclosan-coated suture material has been proven to prevent colonisation by Staphylococcus aureus, multiresistant Staphylococcus aureus, Staphylococcus epidermis, multiresistant Staphylococcus epidermis and Escherichia coli in vitro and in vivo and is therefore suggested to prevent surgical site infection (Rothenburger et al. 2002; Storch et al. 2004; Edmiston et al. 2006; Marco et al. 2007). It is unclear why results of in vitro and in vivo assays were not consistent with the results of the 100 horses evaluated in this study. Possibly, the antibacterial suture material was ineffective against the type and quantity of bacterial isolates in equine midline incisions or equine skin of the actual study population. In this study, the incisions did not undergo routine bacterial culturing; therefore, it is not possible to determine the quantity or type of bacterial isolates present in this population. Before the study period, random incisional cultures were obtained and Staphylococcus spp., Escherichia coli spp. and Enterobacter spp. isolated. These pathogens are similar to the ones isolated in other studies: Streptococcus spp., Staphylococcus spp. and Escherichia coli (Ingle-Fehr et al. 1997), Bacillus spp. and nonhaemolytic Staphylococcus spp. were isolated most commonly after iodophor skin preparation and following wound closure (Galuppo et al. 1999). In vitro and in vivo triclosan was shown to prevent colonisation of most of these pathogens (Rothenburger et al. 2002; Storch et al. 2004; Edmiston et al. 2006; Marco et al. 2007), thus it is unclear why the antibacterial suture material did not show a beneficial effect on equine ventral midline incisions.

Triclosan is a broad-spectrum antiseptic with documented safety and efficacy against selected Gram-positive and Gram-negative bacteria (Jones et al. 2000). It was hypothesised that widespread use of these antibacterial agents may lead to diminished activity against clinically significant pathogens or selection of bacterial strains with increased resistance to commonly used antimicrobial agents (Levy 2001). Selected microbial populations (i.e. Pseudomonas aeruginosa) have been found to be resistant to various antiseptic agents, including triclosan. Other studies (Suller and Russell 2000; Aiello et al. 2004) do not support these findings and suggest that subinhibitory and long-term exposure to triclosan is not associated with increased antimicrobial resistance. In horses it is has been found that potential pathogens, such as Pseudomonas spp., Pasteurella spp., Klebsiella spp. and haemolytic Streptococcus spp., were only isolated sporadically from swab specimens obtained before surgical scrub or after recovery from general anaesthesia. These pathogens were rarely isolated after surgical scrub and skin closure (Galuppo et al. 1999), suggesting that the risk of encountering a clinical strain resistant to triclosan is unlikely in equine ventral midline incisions and therefore unlikely to be the cause of ineffectiveness of antibacterial-coated suture material in equine midline incisions.

It has been shown that there was no significant association between positive wound cultures taken intraoperatively and subsequent drainage/incisional infection (Ingle-Fehr et al. 1997). This fact may minimise the total positive effect that antibacterial-coated suture material can contribute to an equine ventral midline incision, as well as emphasising the important role of other endogenous and exogenous factors in the occurrence of incisional infections. Most equine abdominal surgeries are clean-contaminated emergency surgeries on potentially immunocompromised horses. Heart rate and endotoxaemia at presentation (French et al. 2002), total protein concentration (Mair and Smith 2005a), duration and degree of colic leading to gross contamination of the abdomen prior to presentation (Smith et al. 2007) and duration of surgery (Wilson et al. 1995) have been associated with increased risk for incisional complications. No such associations could be found in this study. Previous studies also found that incisional infection is strongly associated with performing an enterotomy (Honnas and Cohen 1997). Other studies (Kobluk et al. 1989; Phillips and Walmsley 1993; Mair and Smith 2005a), including this study, did not support this finding. Phillips and Walmsley (1993) also reported a higher incidence of wound suppuration in horses with caecum/large colon obstructions. This study did not support this finding. There was no significant association between anatomical location of primary lesion (small intestinal vs. large intestinal) and incisional complications.

Dissection of the linea alba to help placement of the sutures and applying a stent bandage were associated with an increased rate of wound infections (Mair and Smith 2005a). In the present study these procedures were not performed and therefore did not contribute to the incisional infection rate in this study. Antimicrobial administration into the wound during closure (Mair and Smith 2005a) and administration of an iodophore impregnated self-adhesive drape to protect the incision for the recovery have a beneficial effect on wound infection (Galuppo et al. 1999) and were performed on all the horses in this study. However, it was not recorded, in this study, if the self-adhesive drape moved or was pulled during the recovery from general anaesthesia.

Wound healing characteristics, such as healing progress, infection, oedema, erythema, skin temperature, seroma, suture sinus and pain, were analysed in human paediatric patients undergoing various surgical procedures. Significantly fewer patients treated with antibacterial-coated suture material reported pain on Day 1 and there was a trend toward a lower incidence of wound oedema, suggesting that triclosan diminished the inflammatory reaction resulting from subclinical infection (Ford et al. 2005). It is important, however, to emphasise that pain studies in paediatric patients may be subjective, particularly in nonverbal infants. This study does not support the results found by that study. Pain was not significantly different between suture materials. At time point 2 there was a trend toward increased incidence of incisional oedema when triclosan-coated suture material was used (P = 0.081). Overall, 276 horses would be necessary to achieve an 80% power and to detect a significant result when using antibacterial-coated vs. non-antibacterial suture material. This number of horses supports the conclusion that antibacterial-coated suture material is likely to have minimal clinical benefits, as the effect is very small if this many horses are required to achieve a statistically significant difference between groups.

Studies conclude that triclosan has no carcinogenic, genotoxic, pyrogenic or teratogenic effects and systemic levels of triclosan were proven to be extremely low (Barbolt 2002). Recently, however, there has been concern about the formation of toxic by-products of triclosan containing soaps (Fiss et al. 2007). Laboratory studies reinforce that the formation of chloroform and other chlorinated potentially toxic by-products can occur when triclosan containing antimicrobial products react with free chlorine (Fiss et al. 2007). A recent clinical published study in women undergoing breast reduction showed an increased incidence of wound dehiscence when treated with Vicryl Plus (Deliaert et al. 2009). This clinical study concluded that triclosan-coated suture material might have potential adverse effects. The safety of triclosan-coated sutures has been investigated in several studies (Ford et al. 2005). Therefore, it is unlikely that toxic by-products play a significant role in the pathogenesis of insignificantly increased incisional oedema in this study but it has to be mentioned. A recent communication and comment on the study performed by Deliaert et al. (2009) stated that neither the required amount of free chlorine, nor the optimum pH, temperature and ultraviolet radiation are present in the post surgical incision to account for toxic products (Assadian et al. 2009).

Absence of incisional oedema 24–36 h post operatively was associated significantly with a shorter mean incision length. It is unclear what the cause is for this finding. Possibly, a smaller surgical incision may create less inflammation and accompanying oedema. Antibacterial-coated suture material had no significant effect on incisional complications in this horse population and there was a trend toward higher incidence of incisional oedema in the study group compared to the control group. Peri-incisional oedema may affect local tissue oxygen tension at the incisional site and may result in delayed wound healing, suppression of local immune function and provide an optimal environment for bacterial growth (Ingle-Fehr et al. 1997). Therefore, the clinical relevance and benefits of antibacterial-coated suture material in equine ventral midline closure is still questionable.

Manufacturers' addresses

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

2 Orchid Healthcare, Apotex Corp, Weston, Florida, USA.

3 SPSS Inc., Chicago, Illinois, USA.

Author contributions All authors contributed to all aspects of this study.

Ancillary