Six mature horses with large abdominal abscesses (defined as an abscess >15–20 cm in at least one dimension) were treated by surgical drainage and post operative lavage. The abscess was associated with previous intestinal surgery in one horse, and with Streptococcus equi spp. equi infection in the other 5. A Foley catheter was used to drain and lavage the abscess in all cases. The abscess was accessed by a ventral midline laparotomy in 5 cases and by standing flank laparoscopy in one. Two horses were subjected to euthanasia within 7 days due to persistent or recurrent colic. The other 4 horses survived. Lavage of the abscess was continued for a median time of 19 days. Antibacterial therapy was continued until the plasma fibrinogen concentration was normal (median 47 days). Follow-up information was obtained at a median of 1.8 years. All 4 horses were alive at the time of follow-up; 2 horses had suffered one or more bouts of colic that had responded to medical treatment.
Abdominal abscesses can occur in mature horses in a variety of locations and for a variety of reasons, including ulceration, perforation or compromised viability of the gastrointestinal tract, and haematogenous spread of bacteria (including Streptococcus equi spp. equi, Streptococcus equi spp. zooepidemicus and Corynebacterium pseudotuberculosis) to the mesenteric lymph nodes or other abdominal viscera (Rumbaugh et al. 1978; Sweeney et al. 1987; Zicker et al. 1990; Sweeney et al. 1991; Aleman et al. 2003; Elce 2006). Abscesses have been reported in association with the mesenteric lymph nodes, liver, kidney, spleen, uterus, bladder, gastrointestinal tract and abdominal wall, with or without adhesions to multiple structures. Presenting clinical signs can be variable and sometimes vague, but frequently include colic, fever, weight loss, depression and anorexia (Zicker et al. 1990; Elce 2006; Arnold et al. 2010); some horses may have a history of a recent respiratory infection or subcutaneous abscesses.
Treatment options for abdominal abscesses include long-term antimicrobials or surgery (Elce 2006). Depending on the location and involvement of other abdominal structures, surgical options include excision, drainage and lavage, and marsupialisation (Rigg et al. 1987; Prades et al. 1989; Skidell 1996; Elce 2006). Intestinal adhesions associated with abscesses may require further surgical treatment by resection or a by-pass procedure (Prades et al. 1989; Elce 2006). Although some abdominal abscesses may resolve with prolonged antibacterial therapy alone (Aleman et al. 2003; Pusterla et al. 2007), surgical resolution of an abscess is likely to result in a shortened period of antimicrobial administration (Blot and DeWaele 2005; Elce 2006). Previous experience by the authors (unpublished observations) suggests that medical treatment of large abdominal abscesses (i.e. abscesses >15–20 cm in at least one dimension) is usually unsuccessful and surgical treatment is more likely to be effective. However, there are few published reports of the techniques and success rates of surgical treatment of large abdominal abscesses in horses. The purpose of this study was to review the treatment and outcome of a series of 6 mature horses with large abdominal abscesses treated by surgical drainage and post operative lavage.
Details of the 6 horses are summarised in Table 1. Three of the horses (Cases 3, 4 and 6) had been affected by typical S. equi spp. equi infections (strangles) 5 weeks to 3 months previously. One horse (Case 5) had been in contact with other horses affected by strangles 4 weeks previously; this horse developed a left unilateral purulent nasal discharge without other typical signs of strangles. One horse (Case 1) had a history of abdominal surgery (small intestinal resection and side to side jejunocaecostomy) for strangulation of the distal jejunum and proximal ileum 6 years previously; this horse had appeared well in the intervening time period prior to developing clinical signs of recurrent colic and weight loss approximately one month prior to presentation. One horse (Case 2) had been purchased by the current owners 12 months earlier and its previous history was unknown.
Table 1. Details of 5 horses with large abdominal abscesses
The mean duration of illness prior to presentation was 13 days (range 2–40 days). Clinical signs identified prior to admission included depression and inappetence in all horses, chronic and/orrecurrent low-grade colic (5 horses) and intermittent pyrexia (rectal temperature up to 40.2°C) (3 horses). One horse (Case 2) had a history of low-grade abdominal pain for 5 days prior to developing severe colic signs.
Three horses (Cases 1, 2 and 6) demonstrated signs of abdominal pain on admission to the hospital and in Case 2 the pain was severe; distended small intestine was detected by rectal examination and transabdominal ultrasonography in this horse. Three horses demonstrated tachycardia on admission (heart rate 60 beats/min in Cases 2 and 4, and 56 beats/min in Case 5) and 5 horses were depressed. Significant results of haematological, serum biochemical and peritoneal fluid analyses at the time of admission to the hospital are summarised in Table 2. Four horses had leucocytosis and neutrophilia, all were hyperproteinaemic and hyperglobulinaemic, and 2 had hypoalbuminaemia. All 6 horses had hyperfibrinogenaemia and serum amyloid A concentration was elevated in the 2 horses in which it was measured. All 6 horses had elevated total nucleated cell counts and total protein concentrations in the peritoneal fluid. Aerobic cultures of peritoneal fluid were negative in all cases.
Table 2. Selected results of haematological, serum biochemical and peritoneal fluid analyses at the time of hospital admission of the 6 horses with abdominal abscesses
Normal reference range
NP: not performed.
Packed cell volume (%)
White blood cells (WBC) ×109/l
Polymorphonuclear leucocytes ×109/l
Total protein g/l
Serum amyloid A mg/l
Peritoneal fluid WBC ×109/l
Peritoneal fluid total protein g/l
Palpation per rectum identified a painful abdominal mass in all horses. Transabdominal ultrasonography was performed in all horses, and transrectal ultrasonography in 4 horses. The abscesses were usually spherical in shape with a smooth to nodular surface. Ultrasonography typically revealed mixed heterogeneous echogenicity surrounded by a thick echogenic capsule. The abscess was located medial to the caecum (Case 1), in the mid-cranial dorsal abdomen (Cases 2 and 3), the mid-cranial ventral abdomen (Cases 4 and 5), or the caudal left abdomen (Case 6). Endoscopic examination of the upper respiratory tract, including the guttural pouches, in Cases 2–6 revealed no significant abnormalities apart from a single chondroid in the left guttural pouch of Case 5.
Three horses underwent a routine ventral midline laparotomy within 24 h of admission to the hospital. In Case 1, an ovoid-shaped abscess (30 × 14 × 14 cm) was located adjacent to the ileocaecal junction and ileal stump. The jejunocaecostomy was healthy and functional, but there were multiple mature adhesions between several adjacent loops of jejunum and mesojejunum to the ileal stump and abscess capsule; this had resulted in a simple obstruction due to multiple areas of kinking of the jejunum. The proximal jejunum was markedly distended, but was easily decompressed distally into the caecum. Surgical resection or by-pass of the intestine affected by adhesions or sectioning of the adhesions themselves was not considered possible. A 26 French gauge Foley catheter1 (with a metal trocar) was bluntly introduced into the abscess and the balloon inflated with 30 ml of sterile water. The catheter exited the abdomen via a stab incision to the right of the laparotomy wound. In Case 2, a large abscess (25 cm diameter) was located at the root of the mesentery. There were mature adhesions between multiple loops of the jejunum and its associated mesentery and the abscess capsule. One area of adhesion had caused a complete nonstrangulating obstruction of the jejunal lumen; it was possible to cut the mesenteric adhesions in this area to free the obstruction, but the other adhesions were considered inoperable. The proximal intestine was emptied by decompressing the intestinal contents distally to the caecum. A 24 French gauge Foley catheter1 was introduced into the abscess as described in Case 1 and the balloon inflated with 30 ml of sterile water. In Case 6, a multilocular abscess, 20 × 15 × 15 cm, was located in the mesojejunum medial to the spleen. Three loops of jejunum were adherent to the abscess capsule. After packing off the area with absorbent drapes, a stab incision was made into the abscess and 750 ml of thick, creamy pus was evacuated using suction. A 26 French gauge Foley catheter1 was introduced into the abscess through the same incision and the balloon inflated (30 ml); the catheter exited the abdomen via an incision to the left of the laparotomy wound.
Standing laparoscopy was performed via both sides of the abdomen in 3 horses (Cases 3, 4 and 5). In Case 3, the 20 cm diameter abscess capsule was clearly identified via the right side, and a 24 French gauge Foley catheter1 was bluntly introduced into the abscess cavity via one of the lateral instrument portals and the balloon inflated (30 ml). In Cases 4 and 5, laparoscopic examination revealed adhesions of multiple loops of small intestine and small colon to the abscess in the mid to ventral abdomen. No clear view to the abscess capsule could be obtained, so these horses subsequently underwent a ventral midline laparotomy under general anaesthesia. In both cases, the abscess (which measured 30 cm in diameter each) could be accessed via one site ventrally, and 26 French gauge Foley catheters1 were introduced bluntly as previously described to exit the abdomen via stab incisions to the right of the ventral laparotomy wound; the balloons were inflated with 30 ml of sterile water. The intestinal and mesenteric adhesions in these cases were considered inoperable.
A mixed growth of S. equi spp. zooepidemicus and Escherichia coli was cultured from purulent material obtained at surgery in Case 1. S. equi spp. equi was cultured in pure growth from samples obtained from Cases 2–6. Culture of a lavage sample of the left guttural pouch of Case 5 yielded S. equi spp. equi.
In all horses, the catheters were allowed to hang freely from the surgical wounds (fixed to the skin by a purse-string suture). The catheters were kept capped apart from when they were being lavaged. Lavage with sterile 0.9% saline or sterile Hartmann's solution was performed once or twice a day. Sterile solution was infused until moderate resistance was felt and it was then allowed to passively drain; this process was repeated 3 or 4 times on each occasion, followed by infusion of 5 mega units of benzylpenicillin sodium (Crystapen Injection)2 in 20 ml of sterile water and capping of the catheter. The volume of solution infused through the catheters varied from 20–65 ml initially, but within 2–3 days increasing resistance to infusion resulted in a reducing volume such that by 5–7 days volumes of 10–15 ml were typically being used. Purulent material mixed with a small amount of blood typically drained from the catheters for the first 2–3 days but then the fluid became serosanguinous for 3–5 days, and thereafter clear. The horses were maintained on gentamicin sulphate3 (6.6 mg/kg bwt i.v. q. 24 h), procaine benzylpenicillin (Depocillin)4 (22,000 iu/kg bwt i.m. q. 12 h), flunixin meglumine (Finadyne solution)2 (0.5 mg/kg bwt i.v. q. 12 h) for 3–5 days post operatively followed by procaine penicillin (22,000 iu/kg bwt i.m. q. 12 h) alone. Additional post operative treatments (analgesics [flunixin meglumine or phenylbutazone (Equipalazone)4], lidocaine [Lidocainhydrochloride 2%]5 and i.v. fluid therapy) were administered as deemed necessary according to the clinical status of the horse.
Post operatively, Case 1 continued to display signs of intermittent abdominal pain, with mild colic occurring every 12–24 h. These episodes of colic were controlled by the administration of analgesics (phenylbutazone or flunixin meglumine), but in view of the persistent nature of the colic, the owner elected for euthanasia on humane grounds 7 days after the surgery. A limited post mortem examination confirmed the surgical findings, with no additional pathological changes in the abdomen, apart from mild generalised peritonitis. The abscess cavity had shrunk to approximately 50% of its original size and contained serosanguinous fluid.
Case 2 developed signs of moderate colic 28 h after surgery. Small intestinal distension was detected by rectal palpation and transabdominal ultrasonography, and reflux of 12 l of gastric fluid was retrieved by nasogastric intubation. The owners elected not to continue with treatment at this point and euthanasia was performed. A limited post mortem examination revealed fluid distension of the stomach and small intestine, but no further abnormalities apart from those found at surgery. The Foley catheter was in place and there was no observable leakage around the site of puncture of the abscess capsule.
The other 4 horses made satisfactory post operative progress. The horses were hospitalised for variable periods ranging from 5–68 days (median duration of hospitalisation 23 days), during which time repeated haematological examinations, peritoneal fluid evaluations and plasma fibrinogen concentrations were monitored twice a week (these same parameters were also monitored approximately weekly after discharge). The duration of hospitalisation was determined primarily by the owners' willingness and competence to perform the treatments (lavage of the abscess via the Foley catheters and intramuscular injections of procaine penicillin). Plasma fibrinogen values increased in all horses after surgery (maximum recorded values 9.9, 8.2, 8.3 and 7.8 g/l in Cases 3–6, respectively) and then fell reaching the normal range between 42 and 52 days after surgery (median 47 days) (Fig 1); antibacterial therapy was stopped at this point in all horses. Total nucleated cell counts in the peritoneal fluid also increased in all horses after the surgery (maximum recorded values 200, 76, 187 and 110 × 109/l in Cases 3–6, respectively), but fell to the normal range between 55 and 70 days (median 57 days). Three of the horses were switched from intramuscular procaine penicillin to oral potentiated sulphonamides (25 mg sulphadiazine and 5 mg trimethoprim/kg bwt per os q. 24 h) (Norodine granules)6 21–40 days after initiation of treatment because of difficulties experienced administering twice daily intramuscular injections. The Foley catheters were removed when there was no more purulent material draining from them and it had become difficult to lavage fluid into the abscesses. Catheter removal was undertaken at a median of 19 days (range 15–29 days). The catheters were easily removed by cutting the purse string suture at the skin and deflating the balloon, and then pulling them from the abdomen. The wounds were left open and would typically drain peritoneal fluid for 24–48 h before granulating and healing. Omental prolapse was not observed. A mild temporary elevation of plasma fibrinogen concentration and nucleated cell count in the peritoneal fluid was observed in 2 of the horses (Cases 4 and 5) within 3–5 days following removal of the Foley catheters.
A 24 French gauge Foley catheter1 was placed per nasum into the left guttural pouch of Case 5 to allow daily flushing of the pouch with sterile 0.9% saline and acetylcysteine solution (Parvolex)7. This was removed after 15 days when the chondroid had dissolved.
Two episodes of mild colic were observed in Case 4 60 and 64 days after surgery. These episodes responded to symptomatic treatment with analgesics. An episode of severe colic occurred in Case 5 68 days after surgery. This responded to analgesic therapy and starvation for 24 h.
Follow-up information was obtained from the owners of the 4 surviving horses by a telephone questionnaire. All 4 horses were alive at the time of follow-up (median time to follow-up 1.8 years, range 6 months to 5 years). Cases 4, 5 and 6 were in full work, but Case 3 was in minimal work due to chronic lameness.
The clinical signs and clinicopathological features of the 6 horses described herein were similar to those reported in horses with abdominal abscesses previously (Rumbaugh et al. 1978; Sweeney et al. 1987; Zicker et al. 1990; Aleman et al. 2003; Pusterla et al. 2007; Arnold et al. 2010). In all cases the abscess was palpable per rectum and was visualised by either transcutaneous or transrectal ultrasonography. Intestinal and mesenteric adhesions to the abscesses were identified at the time of surgery in all 6 horses, and these had resulted in intestinal obstruction in at least 2 cases, and were considered to be the most likely cause of the episodes of colic in the other horses displaying abdominal pain. Treatment options for extensive mature intra-abdominal adhesions are limited and the adhesions persisted in those horses that survived; these adhesions are considered likely to cause intestinal complications in the future. For these reasons, strategies to prevent the development of large abdominal abscesses after S. equi spp. equi infections (such as monitoring systemic indices of inflammation) should be considered so that animals at high risk of developing abscesses are recognised and treated earlier in the disease process. The abscess was associated with S. equi spp. equi in 5 horses and with previous intestinal surgery in one horse. The prolonged period (6 years) between the intestinal surgery and the development of the abdominal abscess in Case 1 could not be explained. However, the absence of any clinical signs until one month prior to presentation implies that the abscess and associated intestinal adhesions had occurred latterly rather than immediately after the initial surgery for colic. In the other 5 horses metastatic spread of S. equi spp. equi to the mesenteric lymph nodes (i.e. ‘bastard strangles’) was the underlying cause of the abscesses. Complications such as bastard strangles occur in up to 20% of horses affected by strangles and they can have a significant mortality rate (Ford and Lokai 1980; Sweeney et al. 1987; House et al. 2009).
There are few published reports of the prognosis for horses with abdominal abscesses. A recent review of the results of medical treatment in 10 mature horses with abdominal abscesses reported an overall survival rate of 40% (Pusterla et al. 2007). Seven horses survived short term and were treated with antimicrobials for a mean duration of 72 days (range 30–131 days). Follow-up information was available for only 5 of these horses, 4 of which recovered and one was subjected to euthanasia. Post mortem examinations of 4 horses revealed large abdominal abscesses (15–30 cm diameter) in each case; the size of the abscesses in the cases that survived following prolonged antibacterial treatment was not recorded, but it is possible that medical treatment is only likely to be effective in horses with small abscesses. In another large review of abdominal abscesses in mature horses, only 15 of 61 horses (24.6%) recovered, but this included horses treated both medically and surgically, and the sizes of the abscesses were not reported. The recovery rate of 66.7% in the current horses with large abdominal abscesses was therefore considered to be satisfactory.
The appropriate duration of antimicrobial administration in horses with abdominal abscesses is unknown; previously recommended variables to guide decision making include plasma fibrinogen concentrations, peripheral nucleated cell counts, ultrasonographic appearance, nuclear scintigraphic evaluation or laparoscopic visualisation (Elce 2006). Plasma fibrinogen concentration was chosen in theses cases as it is both noninvasive and cost effective, and better reflects the intensity of trauma than peripheral nucleated cell counts (Jacobsen et al. 2009). Fibrinogen is an acute phase protein commonly utilised to both diagnose inflammatory conditions and monitor changes in disease activity in the horse (Andrews et al. 1994), but its disadvantages include the lag phase in response to both initiation and resolution of inflammation (Jacobsen et al. 2009). This delayed response of fibrinogen to clinical changes was considered an advantage in these cases as it supported continued antimicrobial administration beyond resolution of inflammation. However, the prolonged administration of antibiotics to these horses undoubtedly added to the overall costs of treatment, and in view of the fact that purulent material ceased draining from the abscesses within a few days of initiating drainage and lavage, it seems likely that a shorter course of treatment would have been effective.
Whether prolonged antibacterial therapy alone would have yielded similar results in these cases is open to speculation. It is the authors' belief that medical therapy alone would have been unlikely to have resulted in a comparable survival rate. It is our contention that antibacterial therapy alone would be unlikely to sterilise such large abscesses due to problems associated with antibacterial drug penetration (Haggett and Wilson 2008). In addition, we suspect that prolonged inflammation associated with large abscesses undergoing medical therapy would be associated with a greater risk of intra-abdominal adhesions, which increases the risk of colic and other potentially fatal complications. However, until such time as a controlled trial comparing the results of medical therapy alone with surgical plus medical therapy for large abdominal abscesses is undertaken, the merits of these different treatment options cannot be compared. Likewise, further assessment of the required duration of antibacterial therapy following drainage of abdominal abscesses would be helpful to establish the most appropriate treatment protocols.
Conflicts of interest
No conflicts of interest have been declared.
Sources of funding
There were no external sources of funding.
We are grateful to colleagues at Bell Equine Veterinary Clinic for assistance with the care of these cases.
1 Mentor, Porges S.A.S., Le Plessis Robinson, France.