Introduction
A sublumbar abscess is an uncommon lesion in small animals, more frequently encountered in dogs than in cats.1,2,3,4,5,6,7,8 The iliopsoas muscles are most commonly affected, but the epaxial muscles can also be involved, leading to development of a draining tract in the flank region.6,9,10 Clinical signs can be nonspecific (eg, signs of back pain, fever, and reduced appetite). Neurologic deficits may be observed if empyema develops secondary to abscess formation.2,11,12 Sublumbar abscesses in dogs and cats have been attributed to various causes, including wound inoculation, migrating vegetal foreign bodies (MVFBs), diskospondylitis, and ovarian pedicle granulomas after ovariectomy6; some are idiopathic.6,10 Migrating vegetal foreign bodies appear to be the most frequently reported source of infection, depending on the patient’s geographic location.2,6,10
Diagnostic imaging is an essential tool for diagnosis, localization, and identification of an underlying cause for sublumbar abscess before surgical treatment. Computed tomography is commonly used, mainly because it enables assessment of large regions of the body and can be used to detect vegetal foreign bodies.8,13,14 Magnetic resonance imaging is also effective for sublumbar region evaluation and MVFB detection but could be more limited than CT for evaluation of large areas.14,15 Ultrasonography may also allow detection and visualization for foreign body removal, especially in the epaxial muscles, but could be more limited than CT for evaluation of the hypaxial muscles and vertebrae, particularly in large-breed dogs.16,17
Various surgical techniques have been described for treatment of sublumbar abscesses, often with intraoperative ultrasound guidance.6,10,18,19 Prognosis varies, with reported recurrence rates ranging from 0 of 10 (0%) to 3 of 6 (50%).6,10,18,19 Most of those studies aimed at describing new surgical techniques and included small numbers of animals (6 to 22).
The purpose of the retrospective study reported here was to describe CT findings, surgical findings, complications, and outcomes for a large sample of dogs with sublumbar abscesses diagnosed by use of CT and treated surgically at a veterinary referral practice and to identify potential prognostic factors for abscess recurrence in these patients. We hypothesized that the CT findings would correspond with surgical findings and that the absence of an identified source of infection during surgery would be positively associated with abscess recurrence.
Materials and Methods
Case selection
Electronic medical records of Centre Hospitalier Vétérinaire Languedocia were searched to identify dogs that underwent surgery for treatment of a sublumbar abscess between January 1, 2010, and December 31, 2018. The search terms for electronic records were as follows: “sublumbar abscess,” “iliopsoas myositis,” “and lumbar draining tracts.” Only dogs that had a CT examination performed ≤ 1 week prior to surgery and had a detailed description of the surgical procedure in the record were included. Only dogs with records that described the surgical approach elected, any intraoperative complications, whether a source of infection was identified, whether intraoperative ultrasonography was performed, and techniques used for abscess drainage were selected for study inclusion. A detailed report of the CT examination as well as the original images was necessary for case inclusion.
Medical records review
All records were reviewed by 2 individuals (EG and AD-M). Signalment, clinical signs, the duration of clinical signs prior to the referral examination, CBC and plasma biochemical analysis results, CT examination findings, surgical techniques and findings, duration of hospitalization, postoperative treatments, and complications were recorded. Short-term follow-up information (up to discharge from the hospital or the time of suture removal) and duration of antimicrobial treatment after surgery were also obtained from the medical records.
Procedures
All CT examinations were performed with dogs under general anesthesia and positioned in sternal recumbency. Anesthetic premedication consisted of acepromazine (10 μg/kg, IV) or midazolam (0.2 mg/kg, IV), and morphine chlorhydrate or methadone was used for analgesia. Anesthesia was induced with propofol, IV, and maintained with isoflurane in oxygen. Acquisitions were performed for the thoracolumbar vertebral region and, if elected, for the thorax, abdomen, or both. A 4-slice helical CT unit (Brightspeed 4; General Electric) was used. The CT settings were adapted to the region of interest and size of the patient, with a slice thickness ranging from 2.5 to 5.0 mm. Additional acquisitions with 1.25-mm slice thickness were performed for areas of interest when deemed necessary. Imaging was performed before and after an IV injection of iodinated contrast medium (2 mL/kg; 350 mg of I/mL) and recorded.
All CT images were reviewed by 2 board-certified radiologists (LB and CB-T) using dedicated viewing stations equipped with an open-source DICOM viewer (Horos version 3.3.6; Horos Project). A sublumbar abscess was identified if a mass characterized by a peripheral contrast enhancement with non–contrast-enhancing central content was detected within the sublumbar musculature. The extent of muscular infiltration and the presence of an MVFB, diskospondylitis, vertebral osteitis or osteomyelitis, empyema, adenomegaly, and thoracic or abdominal lesions were recorded. Muscular infiltration was defined as abnormal and heterogeneously contrast-enhancing muscle with or without asymmetrical enlargement. An MVFB was suspected or identified when a fusiform hyperattenuating structure, sometimes associated with gas bubbles, was present in the affected tissues; for study purposes, indication of an MVFB or suspected MVFB on CT images in the record was treated as identification of an MVFB on CT examination (Figure 1). Vertebral osteitis was identified if irregular and ill-marginated mineralized proliferation was present at the surface of the vertebrae in contact with the involved muscles. Osteomyelitis was identified if the cortical bone of the vertebrae had evidence of lysis affecting the medullary cavity. Diskospondylitis was identified if lysis of the 2 vertebrae adjacent to an intervertebral disk was detected. Empyema was defined as the presence of contrast-enhancing material within the epidural fat in the vertebral canal. Adenomegaly was defined as an enlargement of the medial iliac and sublumbar lymph nodes > 7 mm in its largest dimension. Muscular lesion length was expressed as a number of vertebrae because of the size difference among dogs. The greatest measured length of the lesion was compared with the length of the body of L2 and rounded to the nearest whole value.
Representative reformatted CT images in a retrospective study to compare CT findings with surgical findings and assess their associations with outcomes of dogs with sublumbar abscesses. A—A sagittal plane image from a 2.5-year-old female American Staffordshire Bull Terrier viewed with a soft tissue window. A fusiform hyperattenuating structure (arrow) with associated gas bubbles (arrowhead) is evident in the sublumbar musculature at the level of L4, consistent with a migrating vegetal foreign body that was removed during surgery. B and C—Transverse (B) and sagittal plane (C) images viewed with a bone window showing evidence of diskospondylitis at the level of L2-3 in a 2-year-old female mixed-breed dog. In panel B, gas bubbles (arrowhead) are evident. D—A transverse plane image viewed with a soft tissue window shows a hyperattenuating structure (arrow) and gas bubbles (arrowhead) in a 2-year-old male mixed-breed dog at the level of L3; a migrating vegetal foreign body was strongly suspected (identified as present on CT for purposes of the study) but was not found during surgery.
Citation: Journal of the American Veterinary Medical Association 259, 11; 10.2460/javma.20.07.0403
Surgeries were performed by 2 surgeons certified by the European College of Veterinary Surgeons as specialists in veterinary surgery (AD-M and SG). If surgery was not performed immediately after the CT examination, subsequent anesthesia was induced and maintained with the previously described protocol. Amoxicillin–clavulanic acid (20 mg/kg, IV) was administered prophylactically. A midline laparotomy, paralumbar surgical approach, or both was elected, depending on CT scan findings and whether a draining tract was present. For the laparotomy approach, an incision was made from the xyphoid process to the umbilicus. This incision was extended caudally if necessary. The sublumbar abscess was identified by palpation or by intraoperative ultrasonography after retraction of organs as needed according to the lesion location. When indicated, ultrasonography was performed by an individual certified by the American College of Veterinary Radiology or the European College of Veterinary Diagnostic Imaging as a radiology specialist (LB or CB-T). The probe used was equipped with a sterile cover. The abscess cavity was identified, and the MVFB (when applicable) was removed with hemostatic forceps after an incision was created with a No. 11 surgical blade. Abscesses were drained, fibrotic tissue was resected, and omentalization was performed.
For the paralumbar approach, an elliptical skin incision was made around the draining tract. Subcutaneous tissues and muscles were dissected until the abscess cavity appeared. The MVFB was removed as described if present. Abscesses were drained, and fibrotic tissue and the draining tract were resected. If the dissection extended to the peritoneal cavity, the site was omentalized.
A passive Penrose drain or active Redon drain was left in place at the surgeon’s discretion. Each surgical site was sampled for bacteriologic culture before abundant flushing. The use of intraoperative ultrasonography, MVFB retrieval, and perioperative complications were recorded. All animals received a broad-spectrum antimicrobial treatment, which was adjusted if needed on the basis of culture results. Postoperative analgesia was provided with morphine chlorhydrate (0.1 to 0.2 mg/kg, IV, q 4 h) or methadone (0.1 to 0.2 mg/kg, IV, q 4 h) for 24 to 72 hours. Analgesics and anti-inflammatories were prescribed for administration at home according to the clinician’s preference.
Minor postoperative complications included all events relative to the surgery that were managed medically or resolved without treatment. Major complications were defined as those that required a second surgery or resulted in the dog’s death or euthanasia.
Long-term follow-up
Telephone interviews were performed with owners or referring veterinarians by 1 individual (EG). The follow-up for each dog was completed ≥ 1 year after the surgical procedure. Questions were asked about wound-related complications, recurrence of clinical signs, and management (medical or surgical) of any abscess recurrence. Any mass or draining tract in the flank region or at the same site in the paralumbar area and any other clinical sign leading to a diagnosis of recurrence by use of ultrasonography or CT were considered recurrences. Time to recurrence was recorded; if multiple recurrences were reported, the time to the first recurrence was taken into account. Outcome was considered as excellent if no recurrence occurred within ≥ 1 year after the procedure, fair if ≥ 1 recurrence was reported but complete healing was achieved with the selected management approach, and poor if recurrence occurred and led to the dog’s death or euthanasia. Dogs for which the outcome could not be categorized as described were excluded from the analyses of outcome and recurrence rate.
Statistical analysis
All categorical data were expressed as proportion and percentage of dogs. For continuous variables, normality was assessed with the Shapiro-Wilk test. The results were reported as the mean ± SD for normally distributed data or median and range for nonnormally distributed data. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of CT for the detection of MVFBs were calculated with 95% CIs. For these analyses, surgical findings were used as the gold standard for MVFB presence or absence.
For dogs with a follow-up time > 12 months and those that had abscess recurrence ≤ 12 months after surgery, multiple variables were tested as potential prognostic factors for the outcome of recurrence (yes vs no). The tested variables included the duration of clinical signs; whether the dog had previously undergone surgery for an abscess in the same anatomic location; presence of a swelling or fistula; presence of osteitis or osteomyelitis, diskospondylitis, and muscular lesion length as assessed with CT; positive versus negative bacteriologic culture results; surgical approach; whether an MVFB was identified surgically; and duration of antimicrobial treatment. Univariable analysis was performed with a Fisher exact or Mann-Whitney test. Logistic regression analysis was then used to identify factors associated with abscess recurrence and determine the relevant ORs. Only variables with values of P < 0.20 in the univariate analysis were included in the multivariable analysis. Statistical software (XLSTAT version 2019.3.1; Addinsoft) was used for all analyses. Values of P < 0.05 were considered significant.
Results
Signalment and clinical history
Fifty-one dogs met the study inclusion criteria. There were 27 males and 24 females. The median age was 3 years (range, 0.8 to 12 years), and mean body weight was 23.9 ± 7.5 kg (52.6 ± 16.5 lb). English Setters (n = 7), mixed-breed dogs (6), German Shorthaired Pointers (4), German Shepherd Dogs (3), Beagles (3), English Springer Spaniels (3), Staffordshire Bull Terriers (3), Brittany Spaniels (3), American Staffordshire Terriers (2), Grand Bleu de Gascogne (2), Griffons (not further specified; 2), Belgian Shepherd Dogs (2), Pointers (not further specified; 2), Porcelaines (2) and 1 of each of the following were included: Berger de Beauce, Boxer, Chihuahua, Golden Retriever, Gordon Setter, Bull Terrier, and White Swiss Shepherd. The most common clinical sign was the presence of a draining tract, swelling in the flank region, or both (29/51 [57%]). Other reported clinical signs were fever (17/51 [33%]), weakness (15 [29%]), evidence of back pain (16 [31%]), paresis (5 [10%]), and coughing (5 [10%]). No dogs had a history of trauma to the paralumbar area or bite wounds. The median duration of clinical signs prior to the referral examination was 2 months (range, 3 days to 36 months). Twenty-four dogs had undergone previous surgeries for a paralumbar abscess, with 17 dogs (33%) undergoing 1 surgery, 4 dogs (8%) undergoing 2 surgeries, and 3 dogs (6%) undergoing 3 surgeries. All surgeries consisted of local exploration and abscess debridement by referring veterinarians.
Results of plasma biochemical analysis were available for 12 dogs. Abnormalities included high total protein concentration (n = 1), high globulin concentration (1), high alkaline phosphatase activity (1), and mild hypoglycemia (1). Thirteen dogs had CBC results available. Abnormalities included mild to marked neutrophilic leukocytosis (n = 6), mild anemia (1), eosinophilia (1), and lymphocytosis (1).
CT findings
On CT examination, an MVFB was identified in 33 of 51 (65%) dogs and gas bubbles were identified for 11 of these dogs. The median length of sublumbar muscle lesions was 3 vertebral lengths (range, 1 to 9 vertebral lengths). Other abnormalities of the lumbar vertebral column included vertebral osteitis or osteomyelitis (25/51 [49%]), diskospondylitis (10 [20%]), spondylosis (5 [10%]), and empyema of the vertebral canal (3 [6%]). Thoracic abnormalities were identified in 10 of the 20 dogs that had thoracic imaging performed. Among these, 5 dogs had multiple lesions. Abnormalities included focal bronchiectasis (7/20 [35%]), focal abnormal enhancement or thickening (or both) of the diaphragmatic crura (6 [30%]), mild pneumothorax, (3 [15%]) and focal pleural thickening (3 [15%]). The most frequently identified abdominal abnormality was medial iliac and lumboaortic lymphadenomegaly (12/51 [24%] dogs). One of each of the following features was also noted: gastric foreign body, right ureteral inflammation (suspected secondary to the abscess), a single nodule on the right adrenal gland, and mild peritoneal effusion.
Surgical procedures
Four surgical procedure types were recorded: laparotomy alone (32/51 [63%]), laparotomy with intraoperative ultrasound guidance (14 [27%]), laparotomy and a local paralumbar approach (2 [4%]), and local paralumbar approach alone (3 [6%]). A passive Penrose drain was left in place for 3 (6%) dogs. One (2%) of the dogs had an active drain placed during surgery with a paralumbar approach. An MVFB was removed from 31 of 51 (61%) dogs. No other source of infection was identified. The CT examination results agreed with the surgical findings for MVFB detection (ie, true positive and true negative results were found for CT) for 39 of 51 (77%) dogs (Table 1). These results corresponded to a sensitivity of 84% and specificity of 65% for the CT examination results, with a PPV of 79% and an NPV of 72% (Table 2). Of the 11 dogs that had gas bubbles identified on CT concurrently with an MVFB, 9 had an MVFB removed. No intraoperative complications were recorded.
Comparison of CT examination and surgical findings related to migrating vegetal foreign bodies (MVFBs) in a retrospective study of 51 dogs that underwent surgery for treatment of a sublumbar abscess between January 1, 2010, and December 31, 2018.
| CT result | Surgical finding | |
|---|---|---|
| MVFB present | MVFB absent | |
| MVFB present | 26 (51%) | 7 (14%) |
| MVFB absent | 5 (10%) | 13 (25%) |
Sensitivity, specificity, and predictive values of CT examination for detection of MVFBs associated with sublumbar abscesses in the 51 dogs in Table 1.
| Sensitivity | Specificity | PPV | NPV | ||||
|---|---|---|---|---|---|---|---|
| TP/TP + FN | (% [95% CI]) | TN/TN + FP | (% [95% CI]) | TP/TP + FP | (% [95% CI]) | TN/TN + FN | (% [95% CI]) |
| 26/31 | 84 (67–93) | 13/20 | 65 (43–81) | 26/33 | 79 (63–93) | 13/18 | 72 (52–93) |
FN = False negative results. FP = False positive results. NPV = Negative predictive value. PPV = Positive predictive value. TN = True negative results. TP = True positive results.
An MVFB was removed from 8 of 14 dogs that had intraoperative ultrasonography performed and from 22 of 37 (59%) dogs for which it was not performed. The proportions of findings with agreement between CT and surgical findings were similar when intraoperative ultrasonography was used (11/14 [79%]) and when it was not used (28/37 [76%]).
Postoperative management
The median hospitalization time was 2 days (range, 1 to 4 days). Postoperative antimicrobial treatment consisted of amoxicillin–clavulanic acid (12.5 to 20 mg/kg, PO, q 12 h) for 43 of 51 (84%) dogs, marbofloxacin (2 to 4 mg/kg, PO, q 24 h) for 4 (8%) dogs, and doxycycline (10 mg/kg, PO, q 24 h) for 1 (2%) of the dogs. Two of 51 (4%) dogs received both amoxicillin–clavulanic acid and marbofloxacin, and 1 (2%) of the dogs received amoxicillin–clavulanic acid, marbofloxacin, and metronidazole. Three of 43 (7%) dogs that were initially treated with amoxicillin–clavulanic acid were subsequently prescribed marbofloxacin on the basis of bacteriologic culture results. The median antimicrobial treatment duration was 28 days (range, 21 to 65 days). Results of bacteriologic culture were available for 49 dogs; the results were negative for 21 (43%) dogs, positive with 1 species identified for 23 (47%) dogs, and positive with 2 species identified for 5 (10%) dogs. Bacterial species included Escherichia coli (n = 8), Streptococcus spp (7), Pasteurella multocida (6), Staphylococcus spp (5), Enterococcus faecalis (4), Kluyvera spp (1), Aerococcus viridians (1), and Serratia marcescens (1). Fifty of 51 (98%) dogs received NSAID treatment for a median of 7 days (range, 5 to 15 days) after surgery, consisting of meloxicam (n = 31), firocoxib (13), or cimicoxib (6). One dog that had been treated with corticosteroids prior to surgery received corticosteroids for 7 days. Tramadol (2 to 4 mg/kg, PO, q 12 h) was prescribed for 6 of 51 (12%) dogs, and gabapentin (10 mg/kg, PO, q 8 h) was prescribed for 4 (8%) dogs.
All dogs survived to hospital discharge. Minor complications were recorded for 6 (10%) dogs ≤ 2 weeks after surgery. These included surgical wound infections in 3 dogs and localized inflammation in 2 dogs that resolved within a few days with topical disinfection; 1 dog had a seroma that was treated with aspiration and a compressive bandage that was applied for 1 week. One major complication was recorded; a dog died as a result of hemoabdomen 3 days after surgery (12 hours after hospital discharge).
Long-term follow-up
The median time to last follow-up for study purposes was 3.6 years (range, 0.3 to 8.7 years). One dog was euthanized 5 months after the surgery because of abscess recurrence, and 1 dog died 3 months after the surgery because of septic peritonitis. These 2 dogs were included in the outcome analysis because of known abscess recurrence. One other dog underwent a hemilaminectomy 3 weeks after surgical treatment for the abscess because of worsening neurologic deficits secondary to empyema (diagnosed during the CT examination in this report). This was not considered a recurrence because the condition had been identified during the first examination, but as no neurologic deficit was present at that time, medical management was elected.
Three dogs were excluded from analysis for the outcome of recurrence: the dog that died because of hemoabdomen, 1 dog that was hit by a car 4 months after surgery without any signs of recurrence at that time, and 1 dog that was euthanized 1 month after surgery because of hemiplegia with loss of deep pain sensation. An empyema had been suspected in this dog during the presurgical CT examination but was not treated surgically at the owner’s request because the neurologic deficits were slight. The neurologic signs worsened during the following weeks, and surgical treatment was suggested but declined by the owners.
Abscesses recurred in 12 of 48 (25%) dogs that were included in the analysis. Three dogs had multiple recurrences. The median time to recurrence was 6 months (range, 1 to 38 months). Recurrent abscesses in 10 dogs were managed surgically: 4 had paralumbar approaches, 5 underwent laparotomy alone, and 1 had laparotomy with intraoperative ultrasound guidance. An MVFB was identified on CT and surgery for 1 dog that did not have an MVFB identified by either method at the time of the first surgery. One dog was managed medically with antimicrobial treatment (cephalexin) when recurrence was identified, and 1 was euthanized. For dogs that had multiple recurrences, another surgical treatment was attempted when the first recurrence was identified, and medical management was provided for the subsequent episodes.
All except 2 of the 48 dogs that were included in the analysis for the outcome of abscess recurrence recovered after treatment (the dog that was euthanized because of recurrence and the dog that died of septic peritonitis). Outcome was considered excellent in 36 of 48 (75%) dogs, fair in 10 (21%), and poor in 2 (4%). Results of the univariable analysis for association with recurrence are summarized (Table 3). Only the presence of diskospondylitis on CT examination was significantly (P = 0.012) associated with abscess recurrence in the multivariable logistic regression analysis. The odds of recurrence for dogs with this finding were 8.4 times those for dogs without this finding (OR, 8.4; 95% CI, 1.8 to 65.1). The recurrence rate was 7 of 29 (24%) for dogs that underwent laparotomy alone, 4 of 14 (28%) for those that had laparotomy with intraoperative ultrasound guidance, 0 of 2 for those that had laparotomy and a local paralumbar approach, and 1 of 3 (33%) for those that had a local paralumbar approach alone. There was no significant (P = 1.0) difference in recurrence rates among procedure types.
Results of univariable analysis for the outcome of abscess recurrence in 48 of the 51 dogs in Table 1.
| Variable | No recurrence (n = 36) | Recurrence (n = 12) | P value |
|---|---|---|---|
| Duration of clinical signs (mo) | 1 (0.1–36) | 4 (0.5–24) | 0.062* |
| Previous surgery (yes vs no) | 16 (44) | 8 (67) | 0.905 |
| Osteitis or osteomyelitis† | 16 (44) | 7 (58) | 0.511 |
| Diskospondylitis† | 3 (8) | 5 (42) | 0.017* |
| Muscular lesion length† (No. of vertebral lengths) | 3 (1–9) | 4 (2–6) | 0.194* |
| Surgical identification of MVFB | 22 (61) | 6 (50) | 0.520 |
| Use of intraoperative ultrasonography | 10 (28) | 4 (33) | 1.0 |
| Positive bacteriologic culture result | 18 (53) | 6 (50) | 0.951 |
| Duration of antimicrobial treatment (d) | 21 (15–60) | 21 (15–85) | 0.527 |
| Presence of a swelling or fistula | 21 (58) | 8 (67) | 0.739 |
Data are reported as median (range) or number (%) of dogs. Three dogs were excluded from this analysis (1 that developed hemoabdomen and died shortly after discharge from the hospital, 1 that was euthanized for worsening neurologic deficits attributed to persistent empyema 1 month after surgery, and 1 that was hit by a car 4 months after surgery and had no signs of recurrence at that time).
Variables included in subsequent logistic regression analysis.
Diagnosed by use of CT.
Discussion
In the present study, CT examination and surgical findings yielded the same result for detecting the presence or absence of MVFBs in 39 of 51 (77%) dogs with sublumbar abscesses, and the only assessed factor that was significantly associated with abscess recurrence in multivariable analysis was the presence of diskospondylitis identified during the CT examination. Our results for the utility of CT to identify MVFBs were consistent with the finding in a previous study9 that 29 of 37 (78%) dogs and cats undergoing surgery for draining tracts had surgical findings that corresponded to preoperative CT results in regard to the presence of a source of inflammation or infection. In 7 of 51 (14%) dogs in our study, an MVFB was identified as present on CT examination but was not found during surgery. This might have been attributable to the presence of fibrotic tissue (which could appear very similar to an MVFB on CT), inadvertent aspiration of the MVFB during surgery if a large amount of purulent liquid was present in the abscess, or localization of the MVFB within fibrotic tissue so that it was not identified during debridement. It is also possible that the MVFB was present but was not identified or removed during surgery. For 1 dog, an MVFB was identified by CT and surgically removed when the abscess recurred, but the source had not been identified during either procedure at the time of the first surgery. No false positive results were reported for CT examinations in previous studies8,13,20 of dogs and cats with nonrespiratory foreign bodies, and the cause for this discrepancy between our results and those in the previous publications was unclear. There might have been a bias toward detection of sublumbar foreign bodies in the present study. Indeed, the prevalence of MVFBs was very high in our region, and we tried to identify a foreign body in as many cases as possible to give the patient the highest chance of recovery. This may have led to overinterpretation in some cases, which would have inevitably lowered the specificity of the CT examination findings for this purpose. In 5 of the 51 (10%) dogs in the present study, an MVFB that was not identified during the CT examination was found during surgery. This was less than previously reported rates of false negative results for CT identification of vegetal foreign bodies in small animals with recurrent draining tracts or subcutaneous vegetal foreign bodies, which range from 10 of 29 (34%) to 14 of 14 (100%).8,9,14 Previous studies14,15,16,17,18,19,20 have shown that vegetal foreign bodies can have a wide range of attenuation values on CT examination, and these are sometimes similar to the values for surrounding soft tissues, making identification more difficult. In our study, when dogs were referred for exploration of recurrent draining tracts or swelling in the paralumbar region, referring veterinarians were asked to stop antimicrobial treatment for ≥ 1 week prior to presentation. This was done to maximize the chance of identifying a large non–contrast-enhancing area during the CT examination and facilitate MVFB localization. Our study results suggested that gas-attenuating bubbles may be a sign that can aid identification of an MFVB on CT images, as 9 of 11 dogs that had this finding had an MFVB removed during surgery. To our knowledge, this potential association has not been previously described; it was not analyzed by statistical methods, and further studies are warranted to investigate the utility of this sign.
Other investigators have shown that the CT attenuation values of wooden foreign bodies may increase with the time spent in the patient,20,21 and this might explain why the only MVFB removed after abscess recurrence was identified during the second CT examination but not the first. Other diagnostic imaging techniques can be valuable for the detection of MFVBs. Magnetic resonance imaging has comparable sensitivity to CT but overall is less accessible in practice.15,22 Moreover, although MRI can be used to scan large body areas, CT may be preferable because MRI is more influenced by artifacts, provides minimal information about thoracic structures, and requires general anesthesia, whereas CT can be performed with heavy sedation.14,15 Ultrasonography is also sensitive17,23,24 and had 100% accuracy for MVFB identification in paralumbar abscesses in 1 study.18 However, its usefulness can be limited by the artifacts generated by surrounding bone structure or gas or by the depth of region to be evaluated in large dogs.25,26 In addition, although it ultrasonography is useful for targeted applications, clipping of large areas of fur may be needed if it is to be used for screening or evaluation of large body regions. It is also poorly sensitive for the detection of osteomyelitis, empyema, diskospondylitis, and lung lobe and diaphragmatic lesions, compared with CT examination.26,27,28 Sinography of draining tracts has shown good results in the past, but its use is limited to a subset of patients with large or long foreign bodies.29,30
The abscess recurrence rate in our study sample (12/48 [25%]) was within the range of those previously reported (0% to 50%).6,10,18,19 Studies by Birettoni et al18 and Morretti et al19 specifically focusing on paralumbar abscesses did not report recurrence over 1 year but focused on a single surgical technique description or patients that had an MFVB effectively removed, and reported recurrence rates of 0%. A possible reason for this difference, compared with our observed recurrence rate, may have been the long-term follow-up pursued for the present study, as this included recurrences up to 38 months after the initial surgery at our facility; for the dog that had the longest recurrence interval, an abscess was detected at the same anatomic location identified previously, and no foreign body or other source of infection was found. Moreover, diskospondylitis was reported in only 1 aforementioned study,18 and it was found in only 1 of 22 dogs. In the present study, diskospondylitis was found in 10 of 51 (20%) dogs and was the only analyzed variable that was significantly associated with abscess recurrence. Diskospondylitis recurrence has been reported even with long-term antimicrobial administration31,32 and could lead to abscess recurrence. Although diskospondylitis seemed to be secondary to paralumbar infection in our study (a MVFB was removed in every case), primary diskospondylitis leading to paralumbar abscess has been suspected previously.33,34 In our study, various surgical techniques and approaches were included; the surgical technique that was considered to provide the greatest visibility for the area of interest and give the best opportunity to remove the MVFB with minimal risk to the patient was chosen for each dog on the basis of CT imaging results. In 1 study9 that included different surgical approaches to recurrent draining tracts depending on whether foreign bodies were identified, suspected, or not identified or suspected during preoperative CT examinations, the recurrence rate was comparable to that in our study (7/37 [19%]).
The abscess recurrence rate was lower for dogs that had an MVFB removed in the present study than for those that did not, similar to previously reported6 findings, even though MVFB identification during surgery was not significantly associated with this outcome. The surgical approach was also not associated with recurrence, but this may have resulted from insufficient statistical power owing to the low number of dogs that underwent each of the 4 surgical approaches. Recurrence rates for the 2 larger groups were similar (7/29 [24%] for laparotomy alone and 4/14 [29%] for laparotomy with intraoperative ultrasonography). However, recent studies focusing on the use of intraoperative ultrasonography in dogs with MVFBs reported no recurrence of abscesses 12 months after surgery, even with a minimally invasive approach, and demonstrated the usefulness of intraoperative ultrasonography for MVFB detection.18,19 In our study, intraoperative ultrasonography was used only for complicated surgeries owing to localization of identified MVFBs on CT examination, lack of MVFB identification on CT examination, or the large size of muscles involved. This may explain why intraoperative ultrasonography use did not influence recurrence in the present study and why the proportions of dogs that had MVFBs removed with and without the technique were similar.
In the study reported here, all dogs survived to hospital discharge, and 50 of 51 (98%) dogs were alive 2 weeks after surgery. The only dog that died during this period developed a hemoabdomen, which might have resulted from local vasculature injury during surgery, most probably involving a vertebral artery because of location of the MVFB. This might have led to delayed vascular wall rupture and hemorrhage. The anatomic location of paralumbar abscesses can require dissection around large vascular structures such as the aorta, caudal vena cava, or vertebral arteries during surgical treatment. In 1 study,10 a dog sustained an aortic rupture during surgery, and the lesion was successfully managed. Great care should then be taken when exploring sublumbar lesions adjacent to the aorta or vertebral arteries.
An MVFB was removed from 30 of 51 (59%) dogs in the present study, which was a higher proportion than in previous studies6,10 of sublumbar and retroperitoneal abscess treatment in dogs. This might have been attributable to regional differences. Indeed, grass awns, spikelets, and seeds are frequently encountered in the south of France and are a frequent source of subcutaneous abscesses or bronchial foreign body leading to pneumothorax.35 Another possible explanation might be differences in the dogs’ lifestyles, as our study sample comprised mainly hunting dogs that could be predisposed to grass awn inhalation during hunting or related exercises.5,7
The migration of vegetal foreign bodies is unidirectional, following low-resistance pathways such as along the respiratory tract or between muscle fasciae. For sublumbar abscesses, it is assumed that MVFBs, after progression inside a bronchus, reach the iliopsoas muscles passing through the diaphragm and its crura. Thoracic abnormalities were found on CT examination for 10 of the 51 dogs in our study (10 of 20 that underwent thoracic imaging), which was consistent with this theory.12,25,35,36 It was also supported by bacteriologic culture results: P multocida, an organism found in the respiratory tract, was detected in 6 dogs. Other studies2,4 have also found bacteria commonly associated with the respiratory tract within sublumbar abscesses. Most of the other bacteria cultured are resident flora from the skin of dogs and likely attributable to the presence of a draining tract or a transcutaneous migration pathway.37
The present study had limitations, some of which were related to its retrospective nature. There was no standardization of CT scan acquisitions, which could have been responsible for diagnosis biases and influenced sensitivity, specificity, NPV, and PPV results. Surgical management, postoperative treatments, or both also differed among clinicians, and this may have influenced outcomes or contributed to the finding that the surgical technique was not associated with abscess recurrence, although further studies are warranted to investigate this. Furthermore, consistent short-term medical recheck examinations were not performed at the referral hospital, which could have lowered the rate of minor postoperative complications. Long-term follow-up was performed through telephone interviews, some of which took place years after the surgery, and some results could have been influenced by bias in the owners’ responses. Computed tomography is very useful for evaluation of deep structures, particularly vertebrae and sublumbar muscles. As its cost in our referring hospital represented only about 20% to 30% of the overall care cost, all owners in this study consented to CT scanning for their dogs prior to surgery. During the retrospective evaluation of records, only 1 dog was excluded because of the lack of preoperative CT examination. However, CT scanning is less accessible for veterinary patients, compared with ultrasonography, and it is unknown whether this method has greater utility for identification of sublumbar abscesses and surgical planning than ultrasonography. Assessment of the usefulness of CT as part of the management of sublumbar abscesses in dogs would require more studies comparing multiple imaging techniques, and this was beyond the scope of our study.
Our results suggested that dogs with sublumbar abscesses have a good prognosis with appropriate surgical planning and treatment. The identification of diskospondylitis on CT examination prior to surgery was associated with significantly increased odds of abscess recurrence, compared with the odds for dogs that did not have this finding, whereas removal of an MVFB during surgery was not associated with this outcome.
Acknowledgments
No third-party funding or support was received in connection with this study or the writing or publication of the manuscript. The authors declare that there were no conflicts of interest.
References
- 1. ↑
Cola V, Del Magno S, Valentini S, et al. Deep vegetal foreign bodies in cats: a retrospective study of 10 cases. J Am Anim Hosp Assoc. 2019;55(5):249–255.
- 2. ↑
Combs M, Decker A, Young P, et al. Grass seed foreign body-related disease in dogs and cats: a wide spectrum of clinical presentations. Aust Vet Pract. 2017;47(1):13–24.
- 3. ↑
Della Santa D, Mannucci T, Busoni G, Citi S. Sublumbar grass awns in two cats: ultrasonographic features and ultrasound-guided retrieval. JFMS Open Rep. 2019;5(2):2055116919885677. doi: 10.1177/2055116919885677
- 4. ↑
Hicks A, Golland D, Heller J, Malik R, Combs M. Epidemiological investigation of grass seed foreign body-related disease in dogs of the Riverina District of rural Australia. Aust Vet J. 2016;94(3):67–75.
- 5. ↑
Brennan KE, Ihrke PJ. Grass awn migration in dogs and cats: a retrospective study of 182 cases. J Am Vet Med Assoc. 1983;182(11):1201–1204.
- 6. ↑
Marvel SJ, MacPhail CM. Retroperitoneal abscesses in seven dogs. J Am Anim Hosp Assoc. 2013;49(6):378–384.
- 7. ↑
Schultz RM, Zwingenberger A. Radiographic, computed tomographic, and ultrasonographic findings with migrating intrathoracic grass awns in dogs and cats. Vet Radiol Ultrasound. 2008;49(3):249–255.
- 8. ↑
Vansteenkiste DP, Lee KCL, Lamb CR. Computed tomographic findings in 44 dogs and 10 cats with grass seed foreign bodies. J Small Anim Pract. 2014;55(11):579–584.
- 9. ↑
Bouabdallah R, Moissonnier P, Delisle F, et al. Use of preoperative computed tomography for surgical treatment of recurrent draining tracts. J Small Anim Pract. 2014;55(2):89–94.
- 10. ↑
Woodbridge N, Martinoli S, Cherubini GB, Caine A, Nelissen P, White R. Omentalisation in the treatment of sublumbar abscessation: long-term outcome in 10 dogs. Vet Rec. 2014;175(24):625.
- 11. ↑
Naughton JF, Tucker RL, Bagley RS. Radiographic diagnosis—paraspinal abscess in a dog. Vet Radiol Ultrasound. 2005;46(1):23–26.
- 12. ↑
Sutton A, May C, Coughlan A. Spinal osteomyelitis and epidural empyema in a dog due to migrating conifer material. Vet Rec. 2010;166(22):693.
- 13. ↑
Jones JC, Ober CP. Computed tomographic diagnosis of nongastrointestinal foreign bodies in dogs. J Am Anim Hosp Assoc. 2007;43(2):99–111.
- 14. ↑
Ober CP, Jones JC, Larson MM, Lanz OI, Werre SR. Comparison of ultrasound, computed tomography, and magnetic resonance imaging in detection of acute wooden foreign bodies in the canine manus. Vet Radiol Ultrasound. 2008;49(5):411–418.
- 15. ↑
Holloway A, Dennis R. McConnell F, Herrtage M. Magnetic resonance imaging features of paraspinal infection in the dog and cat. Vet Radiol Ultrasound. 2009;50(3):285–291.
- 16. ↑
Armbrust LJ, Biller DS, Radlinsky MG, Hoskinson JJ. Ultrasonographic diagnosis of foreign bodies associated with chronic draining tracts and abscesses in dogs. Vet Radiol Ultrasound. 2003;44(1):66–70.
- 17. ↑
Gnudi G, Volta A, Bonazzi M, Gazzola M, Bertoni G. Ultrasonographic features of grass awn migration in the dog. Vet Radiol Ultrasound. 2005;46(5):423–426.
- 18. ↑
Birettoni F, Caivano D, Rishniw M, et al. Preoperative and intraoperative ultrasound aids removal of migrating plant material causing iliopsoas myositis via ventral midline laparotomy: a study of 22 dogs. Acta Vet Scand. 2017;59(1):12. doi: 10.1186/s13028-017-0280-5
- 19. ↑
Moretti G, Birettoni F, Caivano D, et al. Mini-invasive approach for removal of iliopsoas migrating grass awns with an atraumatic wound retractor. J Small Anim Pract. 2021;62(2):150–155.
- 20. ↑
Lamb CR, Pope EHW, Lee KCL. Results of computed tomography in dogs with suspected wooden foreign bodies. Vet Radiol Ultrasound. 2017;58(2):144–150.
- 21. ↑
Imokawa H, Tazawa T, Sugiura N, Oyake D, Yosino K. Penetrating neck injuries involving wooden foreign bodies: the role of MRI and the misinterpretation of CT images. Auris Nasus Larynx. 2003;30(suppl):S145–S147.
- 22. ↑
Young B, Klopp L, Albrecht M, Kraft S. Imaging diagnosis: magnetic resonance imaging of a cervical wooden foreign body in a dog. Vet Radiol Ultrasound. 2004;45(6):538–541.
- 23. ↑
Staudte KL, Hopper BJ, Gibson NR, Read RA. Use of ultrasonography to facilitate surgical removal of non-enteric foreign bodies in 17 dogs. J Small Anim Pract. 2004;45(8):395–400.
- 24. ↑
Attanasi G, Laganga P, Rossi F, et al. Utilizzo dell'ecografia e della TC nella diagnosi e nel trattamento dei corpi estranei vegetali in 56 cani. Veterinaria. 2011;25(1):25–30.
- 25. ↑
Frendin J, Funkquist B, Hansson K, Lönnemark M, Carlsten J. Diagnostic imaging of foreign body reactions in dogs with diffuse back pain. J Small Anim Pract. 1999;40(6):278–285.
- 26. ↑
Kirberger RM. Early diagnostic imaging findings in juvenile dogs with presumed diskospondylitis: 10 cases (2008–2014). J Am Vet Med Assoc. 2016;249(5);539–546.
- 27. ↑
Ruoff CM, Kervin SC, Taylor AR. Diagnostic imaging of discospondylitis. Vet Clin Small Anim Pract. 2018;48(1);85–94.
- 28. ↑
Cavaino D, Birettoni F, Rishniw M, et al. Ultrasonographic findings and outcomes of dogs with suspected migrating intrathoracic grass awns: 43 cases (2010–2013). J Am Vet Med Assoc. 2016;248(4);413–420.
- 29. ↑
McEvoy FJ, Lamb CR, White RN. An application of sinography in small animal practice. Vet Rec. 1993;132(8):183–185.
- 30. ↑
Lamb CR, White RN, McEvoy FJ. Sinography in the investigation of draining tracts in small animals: retrospective review of 25 cases. Vet Surg. 1994;23(2):129–134.
- 31. ↑
Kraft SL, Mussman JM, Smith T, Biller DS, Hoskinson JJ. Magnetic resonance imaging of presumptive lumbosacral discospondylitis in a dog. Vet Radiol Ultrasound. 1998;39(1):9–13.
- 32. ↑
Schwartz M, Boettcher IC, Kramer S, Tipold A. Two dogs with iatrogenic discospondylitis caused by meticillin-resistant Staphylococcus aureus. J Small Anim Pract. 2009;50(4):201–205.
- 33. ↑
Johnston DE, Christie BA. The retroperitoneum in dogs: retroperitoneal infections. Compend Contin Educ Pract Vet. 1990;12(8):1035–1045.
- 34. ↑
Packer RA, Coates JR, Cook CR, Lattimer JC, O'Brien DP. Sublumbar abscess and diskospondylitis in a cat. Vet Radiol Ultrasound. 2005;46(5):396–399.
- 35. ↑
Baudin Tréhiou C, Gibert S, Sériot P, Dunié Mérigot A, Blond L. CT is helpful for the detection and presurgical planning of lung perforation in dogs with spontaneous pneumothorax induced by grass awn migration: 22 cases. Vet Radiol Ultrasound. 2020;61(2):157–166.
- 36. ↑
Hopper BJ, Lester NV, Irwin PJ, Eger CE, Richardson JL. Imaging diagnosis: pneumothorax and focal peritonitis in a dog due to migration of an inhaled grass awn. Vet Radiol Ultrasound. 2004;45(2):136–138.
- 37. ↑
Grösslinger K, Lorinson D, Hittmair K, Konar M, Weissenböck H. Iliopsoas abscess with iliac and femoral vein thrombosis in an adult Siberian husky. J Small Anim Pract.2004; 45(2):113–116.
