Association between preoperative characteristics and risk of anaesthesia-related death in dogs in small-animal referral hospitals in Japan.

OBJECTIVE: To explore the major risk factors linking preoperative characteristics and anaesthesia-related death in dogs in referral hospitals in Japan. STUDY DESIGN: Observational cohort study. ANIMALS: From April 1, 2010 to March 31, 2011, 4323 dogs anaesthetized in 18 referral hospitals in Japan. METHODS: Questionnaire forms were collated anonymously. Death occurring within 48 hours after extubation was considered as an anaesthesia-related death. Patient outcome (alive or dead) was set as the outcome variable. Preoperative general physical characteristics, complete blood cell counts, serum biochemical examinations and intraoperative complications were set as explanatory variables. The risk factors for anaesthesia-related death were evaluated using chi-square test or Fisher's exact test, followed by multivariable logistic regression analysis of the data. Significance was set at p < 0.05. RESULTS: Thirteen dogs that died from surgical error or euthanasia were excluded from statistical analysis. The total mortality rate in this study was 0.65% [28/4310 dogs; 95% confidence interval (CI), 0.41-0.89]. Furthermore, 75% (95% CI, 55.1-89.3) of anaesthesia-related deaths occurred in dogs with pre-existing diseases. Most of the deaths occurred postoperatively (23/28; 82.1%; 95% CI, 63.1-93.9). Preoperative serum glucose concentration <77 mg dL-1 (6/46; 13.0%; 95% CI, 4.9-26.3), disturbance of consciousness (6/50; 12.0%; 95% CI, 4.5-24.3), white cell count >15,200 µL-1 (16/499; 3.4%; 95% CI, 1.9-5.5) and American Society of Anesthesiologists grade III-V (19/1092; 1.7%; 95% CI, 1.1-2.7) were identified as risk factors for anaesthesia-related death. Intraoperative hypoxaemia (8/34; 23.5%; 95% CI, 10.7-41.2) and tachycardia (4/148; 2.7%; 95% CI, 0.7-6.8) were also risk factors for anaesthesia-related death. CONCLUSIONS AND CLINICAL RELEVANCE: The results revealed that certain preoperative characteristics were associated with increased odds of anaesthesia-related death, specifically low serum glucose concentration and disturbances of consciousness. Greater attention to correcting preanaesthetic patient abnormalities may reduce the risk of anaesthesia-related death.

Laparoscopic treatment of congenital portosystemic shunts with portal pressure measurement and portal angiography in 36 dogs.

Introduction: Laparoscopic surgery is used for canine congenital extrahepatic portosystemic shunts (CEHPSS). However, outcomes of laparoscopic surgery involving simultaneous portal vein angiography and portal pressure measurement to attenuate or completely occlude the shunt vessel in canines remain unclear. This study aimed to evaluate outcomes and complications of laparoscopic portosystemic shunt occlusion (LAPSSO) for CEHPSS. Methods: Between June 2014 and March 2021, data on dogs undergoing cellophane banding (CB) and complete occlusion of laparoscopically treated congenital extrahepatic port shunts were collected from hospital records. Cases in which complete occlusion was laparoscopically performed, or a CB was used for gradual occlusion were included. A total of 36 dogs (14 males; median age 32.5 months [range, 5-99] with median body weight, 4.2 kg [range, 1.5-7.9]) that underwent LAPSSO for CEHPSS were included. All the dogs underwent computed tomographic angiography (CTA), and data on blood and radiological examinations were collected. Shunt vessel morphology was categorized using CTA findings. Portal pressure measurements and portal angiography were performed by accessing mesenteric and splenic veins in 30 and 6 cases, respectively. Results: The most common shunt types were spleno-phrenic shunts 16/36 (44.4%), followed by spleno-azygos 9/36 (25.0%), spleno-caval 4/36 (11.1%), right gastric-caval 6/36 (16.6%), and right gastric-caval with caudal loop shunts 1/36 (2.7%). The median portal pressure after complete occlusion was 11.5 mmHg (range, 4-16); portal pressures in the two dogs undergoing CB attenuation were 22 and 24 mmHg. The median operating time in the dogs with right (n = 25) and left (n = 11) recumbent positioning was 55 min (range, 28-120) and 54 min (range, 28-88), respectively. One dog had pneumothorax due to injury to the diaphragm. Another dog developed postoperative hypernatremia and succumbed 5 h post-procedure. Nevertheless, no other dogs exhibited signs of portal hypertension within 72 h. Blood tests and abdominal ultrasounds performed 1-2 months postoperatively revealed no residual shunts. Discussion: LAPSSO, coupled with portal pressure measurement and portal angiography, was shown as safe and effective approach that facilitated successful occlusion of CEHPSS. Further large-scale prospective studies and analyses of perioperative complications are needed.

Dural Changes Induced by an Ultrasonic Bone Curette in an Excised Porcine Spinal Cord.

In spinal surgery, ultrasonic bone curettes are considered unlikely to cause mechanical injury to the dura; however, there is little evidence to support this claim. We investigated the effect of direct contact with an ultrasonic bone curette on the dura and the protective effect of covering the dura with a cotton pattie using an excised porcine spinal cord. The ultrasonic bone curette was pressed against the porcine spinal cord with constant force and activated for 1 s, with or without covering the dura with a cotton pattie. The dural surface and cross-section were observed using electron and light microscopy. When the ultrasonic bone curette was applied directly against the dura, most specimens showed non-perforating dural injuries. However, none of the specimens showed dural perforation. Histological changes were also observed. The use of a cotton pattie reduced the occurrence of these changes, although it did not prevent them when ultrasonic vibration was applied with a large force. We considered ultrasonic bone curettes to have a low risk of dural perforation and, thus, to be a safe surgical device as long as they did not accidentally make strong contact with the dura.

Identification of domestic cat hepadnavirus from a cat blood sample in Japan.

The hepatitis B virus (Hepadnaviridae) induces chronic hepatitis and hepatic cancer in humans. A novel domestic cat hepadnavirus (DCH) was recently identified in several countries, however, the DCH infection status of cats in Japan is unknown. Therefore, we investigated the DCH infection rate of 139 cat samples collected in Japan. We identified one positive blood sample (0.78%) from a 17-year-old female cat with chronically elevated alanine aminotransferase. Phylogenetic analysis demonstrated that the DCH strain identified in this study is genetically different from strains in other countries. Further investigations are required to elucidate the evolution of DCH and the impact of DCH infection on hepatic diseases in domestic cats.

Pancreatic enzyme activity in the bile of healthy cats and its association with biliary morphology.

BACKGROUND: In human medicine, congenital maljunction of the common bile duct (CBD) and main pancreatic duct (MPD), or pancreatobiliary maljunction (PBM), is a known cause of cholecystitis. OBJECTIVE: Pancreatic enzyme activity in the bile (a diagnostic marker for PBM) of healthy cats was measured to determine normal values and evaluate its relationship with biliary morphology. ANIMALS: Fifty-two healthy cats. METHODS: Cross-sectional study of the biliary tracts of healthy cats during laparoscopic ovariohysterectomy and measurement of pancreatic enzyme activity in bile. The cats were divided into groups A and B based on the ratio of the diameter of the cystic duct (CD) to the CBD. The normal ratio was 3.4. Pancreatic enzyme activity in bile was compared between the groups. RESULTS: The CBDs were straight in all cases, whereas the CDs were variably tortuous or dilated. Amylase activity in the bile (median, <100 U/L; range, <100-591 U/L) was lower than in serum in all cases, and group B, which had a CD/CBD ratio >3.4, had significantly higher amylase activity (median, 109 U/L; range, <100-591 U/L) in the bile than did group A (median, <100 U/L; range, <100-238 U/L), which had a CD/CBD ratio <3.4 (P = .0009). CONCLUSIONS AND CLINICAL IMPORTANCE: The results suggest that a dilated CD is associated with reflux of pancreatic juice. In the future, it will be necessary to examine the clinical usefulness of these findings by measuring pancreatic enzyme activity in the bile of cats with cholangitis.

Effects of a single-bolus bupivacaine injection into the coccygeal spinal canal of rabbits.

It has been reported that drugs intended for epidural administration through the lumbosacral junction are accidentally administered into the subarachnoid space frequently in rabbits. Therefore, we evaluated the epidural single-bolus injection technique for the administration of bupivacaine into the coccygeal spinal canal of rabbits. After epidural distribution was confirmed by the injection of iohexol into the coccygeal spinal canal, 0.3 ml/kg 0.5% bupivacaine or 0.3 ml/kg normal saline was injected via the same needle. After the first attempt of iohexol injection, although the contrast was found in the epidural space in all rabbits, the additional contrast was also found in blood vessel in 3 rabbits and in muscular layer in 1 rabbit. Subarachnoid distribution was not observed in any of the rabbits. The time taken to regain normal anal reflex, movement of the hind limbs during walking, conscious proprioception of the hind limbs, and pain sensation of the tail and left hind limb, following coccygeal spinal canal injection, were significantly longer in the bupivacaine group than in the normal saline group. These findings indicated that coccygeal epidural injection of bupivacaine in rabbits may provide anesthesia for the hind limbs, perineum, and tail, but inadvertent vascular entry of the epidural drug may occur.

Usefulness of Urinary Creatinine/Urea Nitrogen Ratio as Indicator of Body Protein Catabolism in Dogs Fed Low Protein Diets.

Low protein diets (LPs) constitute a reportedly effective form of nutritional therapy for canine chronic kidney disease and cirrhosis. These diets have long been feared to result in reduced muscle mass due to protein catabolism. This adverse effect, however, remains largely unrecognized in veterinary medicine as there are no easily applicable catabolism indicators. Therefore, we focused on urinary creatinine, a metabolite of protein in the urine, and examined whether its ratio to urinary urea nitrogen (UCrn/UN) can be used to assess protein catabolism. In Experiment 1, we first consecutively fed seven healthy beagles an LP, standard protein (SP), and high protein (HP) diet for 1 week each and then measured the UCrn/UN ratio at 2-h intervals from fasting to 16 h post-prandially. We consequently found that the UCrn/UN ratio was significantly elevated in the LP pre-prandially and at all post-prandial measurement points (P < 0.01). No significant differences were observed between the SP and HP. Analysis of fasting plasma amino-acid concentrations revealed that the concentration of methionine was significantly lower in the LP than in the other diets (P < 0.05). Although the effects of this change in amino-acid concentration were unclear, the UCrn/UN ratio was considered having increased due to a deficiency in protein and/or amino acids during LP feeding. In Experiment 2, we continuously fed five healthy beagles an LP for 18 weeks and then measured the UCrn/UN ratio as described above. We also measured changes in body composition with computed tomography. At weeks 10 and 18, the fasting UCrn/UN ratio was significantly higher than it was prior to the start of the LP; however, post-prandially, the UCrn/UN ratio decreased to the point that the significant difference disappeared. Muscle mass decreased at weeks 10 and 18. These results suggest that the fasting UCrn/UN ratio could be used as an indicator of protein catabolism in LP feeding. Our experiments thus indicate that examination of potential increases in the UCrn/UN ratio 1 week after introduction of LP feeding to healthy dogs could enable detection of body protein catabolism in long-term feeding of LP before muscle breakdown occurs.

Purification and characterization of feline ghrelin and its possible role.

Ghrelin, a novel 28-amino acid peptide with an n-octanoyl modification at Ser3, has been isolated from rat and human stomach as an endogenous ligand for the growth hormone secretagogue receptor. Here, we purified feline ghrelin and examined its possible physiological role in cats. The major active form of feline ghrelin is a 28-amino acid peptide octanoylated (C8:0) at Ser3; except for one amino acid residue replacement, this structure is identical to those of rat and human ghrelins. However, much structural divergence in peptide length and fatty acid modification was observed in feline ghrelin: peptides consisting of 27 or 26 amino acids lacking Gln14 and/or Arg28 were found, and the third serine residue was modified by octanoic acid (C8:0), decanoic acid (10:0), or unsaturated fatty acids (C8:1, C10:1 and C10:2). In agreement with the structural divergence, two kinds of cDNA with different lengths were isolated. Administration of synthetic rat ghrelin increased plasma growth hormone levels in cats, with a potency similar to that in rat or human. Plasma levels of ghrelin in cats increased approximately 2.5-fold after fasting. The present study indicates the existence of structural divergence in feline ghrelin and suggests that, as in other animals, ghrelin may play important roles in GH release and feeding in cats.

Retrospective analysis of canine gallbladder contents in biliary sludge and gallbladder mucoceles.

The pathophysiology of canine gallbladder diseases, including biliary sludge, gallbladder mucoceles and gallstones, is poorly understood. This study aimed to evaluate the component of gallbladder contents and bacterial infection of the gallbladder in order to elucidate the pathophysiology of biliary sludge and gallbladder mucoceles. A total of 43 samples of canine gallbladder contents (biliary sludge, 21 and gallbladder mucoceles, 22) were subjected to component analysis by infrared spectroscopy, and the resultant infrared spectra were compared with that of swine mucin. Of the 43 samples, 41 were also evaluated by aerobic and anaerobic bacterial culture. The contents of 20 (95.2%) biliary sludge and 22 (100%) gallbladder mucocele samples exhibited similar infrared spectra as swine mucin. Although biliary sludge and gallbladder mucocele contents exhibited similar infrared spectra, one sample of biliary sludge (4.8%) was determined to be composed of proteins. The rate of bacterial infection of the gallbladder was 10.0% for biliary sludge and 14.3% for gallbladder mucoceles. Almost all of the identified bacterial species were intestinal flora. These results indicate that the principal components of gallbladder contents in both gallbladder mucoceles and biliary sludge are mucins and that both pathophysiologies exhibit low rates of bacterial infection of the gallbladder. Therefore, it is possible that gallbladder mucoceles and biliary sludge have the same pathophysiology, and, rather than being independent diseases, they could possibly represent a continuous disease. Thus, biliary sludge could be considered as the stage preceding the appearance of gallbladder mucoceles.

A case report of suspected hepatopulmonary syndrome secondary to ductal plate malformation with chronic active hepatitis in a dog.

Hepatopulmonary syndrome (HPS) is a respiratory complication of hepatic disease, that is well recognized in humans and defined by the presence of 1) liver disease, 2) hypoxemia and/or high alveolar-arterial oxygen gradient (AaDO2) and 3) intrapulmonary vasodilatation. The present report describes a similar case of HPS in a dog. A six-month-old Papillon was diagnosed with ductal plate malformation with chronic active hepatitis and showed progressive increases in AaDO2 over the course of the following six months. The presence of intrapulmonary vasodilatation was confirmed by agitated saline contrast transthoracic echocardiography. Also, the absence of congenital cardiac defect was confirmed by transthoracic echocardiography. From these results, we suspected that this dog had HPS. This is the first description of suspected canine HPS.

Arterial blood gas anomaly in canine hepatobiliary disease.

Arterial blood gas analysis is an important diagnostic and monitoring tool for respiratory abnormalities. In human medicine, lung complications often occur as a result of liver disease. Although pulmonary complications of liver disease have not been reported in dogs, we have frequently encountered hypoxemia in dogs with liver disorders, especially extrahepatic biliary obstruction. In addition, respiratory disorders account for 20% of perioperative fatalities in dogs. Therefore, in this study, we evaluated the respiratory status in dogs with hepatobiliary disease by arterial blood gas analysis. PaO2 and PaCO2 were measured. Alveolar-arterial oxygen difference (AaDO2), the indicator of gas exchange efficiency, was calculated. Compared to healthy dogs (control group), hepatobiliary disease dogs had significantly lower PaO2 and higher AaDO2. Hypoxemia (PaO2 of ≤80 mmHg) was observed in 28/71 dogs with hepatobiliary disease. AaDO2 was higher (≥30 mmHg) than the control group range (11.6 to 26.4 mmHg) in 32/71 hepatobiliary disease dogs. By classifying type of hepatobiliary disease, dogs with extrahepatic biliary obstruction and chronic hepatitis showed significantly lower PaO2 and higher AaDO2 than in a control group. Dogs with chronic hepatitis also had significantly lower PaCO2. The present study shows that dogs with hepatobiliary disease have respiratory abnormalities more than healthy dogs. Preanesthetic or routine arterial blood gas analysis is likely beneficial to detect the respiratory abnormalities in dogs with hepatobiliary disease, especially extrahepatic biliary obstruction and chronic hepatitis.

Tension pneumoperitoneum due to gastric perforation in a cat.

A 14-year-old spayed female cat weighing 2.3 kg developed tension pneumoperitoneum that progressed for 6 hr. Abdominocentesis was performed and 1.3 L of gas was aspirated after which the cat's respiratory condition improved. The next day, the cat's abdomen redistended, and laparotomy was performed. A gastric perforation was noted and sutured after the margin was resection. The cat showed a good physical condition without any gastrointestinal signs at 6 months after surgery.

Relationship between growth and plasma concentrations of ghrelin and growth hormone in juvenile beagle dogs.

Although the release of growth hormone (GH) is known to be regulated mainly by GH-releasing hormone (GHRH) and somatostatin (SRIF) secreted from the hypothalamus, ghrelin also may be involved in GH release during juvenile period. We have examined plasma concentrations of acylated ghrelin, desacyl ghrelin, and GH in juvenile beagle dogs. Plasma acylated and desacyl ghrelin levels changed through aging; however, there was no closely correlation between ghrelin, body weight and circulating GH levels during juvenile period. The increase in body weight was essentially linear until 8 months of age, whereas plasma GH concentrations exhibited bimodal peaks for the meanwhile. The results suggest that ghrelin may not play internal cueing in GH secretion in juvenile beagle dogs.

Laparoscopic cryptorchidectomy in standing bulls.

Laparoscopic cryptorchidectomy without insufflation was applied in 10 standing bulls aged 3 to 15 months. Nine bulls were preoperatively pointed out intra-abdominal testes by computed tomography. Preoperative fasting for a minimum of 24 hr provided laparoscopic visualization of intra-abdominal area from the kidney to the inguinal region. Surgical procedure was interrupted by intra-abdominal fat and testis size. It took 0.6 to 1.5 hr in 4 animals weighing 98 to 139 kg, 0.8 to 2.8 hr in 4 animals weighing 170 to 187 kg, and 3 and 4 hr in 2 animals weighing 244 and 300 kg to complete the cryptorchidectomy. In conclusion, standing gasless laparoscopic cryptorchidectomy seems to be most suitable for bulls weighing from 100 to 180 kg.

Histopathological and enzyme histochemical observations on mast cells in pulmonary arterial lesion of dogs with Dirofilaria immitis infestation.

Histopathological and enzyme histochemical observations were performed on mast cells in pulmonary arterial lesion of Dirofilaria immitis in dogs. The results showed that chymase- and tryptase-positive mast cells were diffusely present in the lesions, especially in the adventitia and proliferated intima. At 2 weeks after surgical worm transplantation, mast cells already appeared in the intima and media, and chymase-positive cells were dominant in the adventitia. Results of this study suggested a possibility that mast cells would be involved in the pathogenesis of pulmonary arterial lesion of dogs with Dirofilaria immitis infestation.

Cardiorespiratory effects of epidural administration of morphine and fentanyl in dogs anesthetized with sevoflurane.

OBJECTIVE: To determine the cardiorespiratory effects of epidural administration of morphine alone and in combination with fentanyl in dogs anesthetized with sevoflurane. DESIGN: Prospective study. ANIMALS: 6 dogs. PROCEDURE: Dogs were anesthetized with sevoflurane and allowed to breathe spontaneously. After a stable plane of anesthesia was achieved, morphine (0.1 mg/kg [0.045 mg/lb]) or a combination of morphine and fentanyl (10 microg/kg [4.5 microg/lb]) was administered through an epidural catheter, the tip of which was positioned at the level of L6 or L7. Cardiorespiratory variables were measured for 90 minutes. RESULTS: Epidural administration of morphine alone did not cause any significant changes in cardiorespiratory measurements. However, epidural administration of morphine and fentanyl induced significant decreases in diastolic and mean arterial blood pressures and total peripheral resistance. Stroke volume was unchanged, PaCO2 was significantly increased, and arterial pH and base excess were significantly decreased. Heart rate was significantly lower after epidural administration of morphine and fentanyl than after administration of morphine alone. None of the dogs had any evidence of urine retention, vomiting, or pruritus after recovery from anesthesia. CONCLUSIONS AND CLINICAL RELEVANCE: Results suggest that epidural administration of morphine at a dose of 0.1 mg/kg in combination with fentanyl at a dose of 10 microg/kg can cause cardiorespiratory depression in dogs anesthetized with sevoflurane.

Disruption of the posterior chamber-anterior hyaloid membrane barrier during phacoemulsification and aspiration as revealed by contrast-enhanced magnetic resonance imaging.

OBJECTIVE: To determine the effect of phacoemulsification and aspiration (PEA) in different surgical settings on the posterior chamber-anterior hyaloid membrane (AHM) barrier. METHODS: Eighty enucleated porcine eyes divided into 8 groups underwent lens extraction at a bottle height of 35 or 95 cm by 1 of the following 4 procedures: standard PEA, standard PEA without hydrodissection (sPEA-), sPEA- including 5 minutes of irrigation of the anterior chamber, and sPEA- including 20 cycles of deflation (5 seconds each cycle) and inflation (10 seconds each cycle) of the anterior chamber. Distribution of gadopentetate dimeglumine (gadolinium-diethylenetriamine pentaacetic acid [Gd-DTPA]) in the irrigating fluid was assessed by magnetic resonance imaging (MRI). RESULTS: The most common MRI pattern was the posterior chamber type, where Gd-DTPA was localized in the anterior and posterior chambers, followed by AHM detachment, in which Gd-DTPA was evident beneath the posterior lens capsule. The least common was AHM tear, in which Gd-DTPA entered the vitreous cavity through a tear in the AHM. Logistic regression analysis revealed prolonged irrigation (P< .001) and deflation/inflation of the anterior chamber (P< .001) as risk factors for AHM detachment and hydrodissection (P= .04) as a risk factor for AHM tear. CONCLUSION: Changes in intraocular pressure can disrupt the posterior chamber-AHM barrier during PEA. CLINICAL RELEVANCE: Cataract surgeons should reexamine their surgical settings to avoid unnecessary stress on the eye.

The effect of body position on the 'hanging drop' method for identifying the extradural space in anaesthetized dogs.

OBJECTIVES: To assess the accuracy of the 'hanging drop method' for identifying the extradural space in anaesthetized dogs positioned in sternal or lateral recumbency. STUDY DESIGN: Prospective randomized-experimental study. ANIMALS: Seventeen clinically healthy adult dogs, 10 females and seven males weighing 8.4-26.2 kg. METHODS: Dogs were positioned in either sternal (n = 8) or lateral (n = 9) recumbency under general anaesthesia. A 20 SWG spinal needle pre-filled with 0.9% saline was advanced through the skin into the lumbosacral extradural space and the response of the saline drop recorded, i.e. whether it: 1) was aspirated from the hub into the needle; 2) remained within the hub, or 3) moved synchronously with i) spontaneous respiration, ii) heart beat or iii) manual lung inflation. The position of the needle tip was ultimately determined by positive contrast radiography. RESULTS: One dog positioned in lateral recumbency was excluded from the study because bleeding occurred from the needle hub. Saline was aspirated into the needle in seven of eight dogs held in sternal recumbency but in none of the dogs positioned in lateral recumbency. Accurate needle tip placement in the extradural space was confirmed by positive contrast radiography in all dogs. CONCLUSION AND CLINICAL RELEVANCE: The 'hanging drop' method, when performed with a spinal needle, appears to be a useful technique for identifying the location of the extradural space in anaesthetized medium-sized dogs positioned in sternal, but not in lateral recumbency. The technique may yield 'false negative' results when performed in dogs positioned in sternal recumbency.

Lymphography of the thoracic duct by percutaneous injection of iohexol into the popliteal lymph node of dogs: experimental study and clinical application.

OBJECTIVE: To evaluate the efficacy of percutaneous administration of iohexol into the popliteal lymph node as a non-invasive technique for thoracic duct lymphangiography in dogs. STUDY DESIGN: Experimental study and clinical report. ANIMALS: Normal adult dogs (n=4) and 1 dog with recurrent chylothorax. METHODS: For the experimental study, 4 dogs (weight, 8.4-12.3 kg) had 5-10 mL iohexol injected percutaneously into 1 popliteal lymph node and then thoracic radiographs were taken. Popliteal lymph nodes were examined by histopathology 8 days later. One 25-kg dog with recurrent chylothorax had 25 mL iohexol injected into the right popliteal lymph node followed by thoracic radiography. RESULTS: In experimental dogs, the thoracic duct was best visualized on thoracic radiographs after administration of 10 mL iohexol. Clinically, no abnormalities were identified in the injected limb and except for 1 dog that had large numbers of siderocytes and erythrophagocytic macrophages in the injected lymph node, the histopathologic findings in the other injected popliteal lymph nodes were not different from contralateral nodes. In the clinical case, the thoracic duct was visualized, but there was leakage of iohexol around the node. CONCLUSION: The thoracic duct in dogs can be visualized by lymphography after percutaneous injection of iohexol (1 mL/kg at 2 mL/min) into the popliteal lymph node. CLINICAL RELEVANCE: Percutaneous popliteal lymph node administration of iohexol should be considered as an alternative to mesenteric lymph node injection for radiographic identification of the thoracic duct in dogs.