Ischaemic distal limb necrosis and Klebsiella pneumoniae infection associated with arterial catheterisation in a cat

Abstract

This case report describes dorsal pedal arterial thrombosis and infection with Klebsiella pneumoniae subsequent to arterial catheter placement in a cat. The complication led to avascular necrosis of the metatarsal and pedal soft tissue. The catheter was placed for blood pressure monitoring during surgery for correction of a peritoneopericardial diaphragmatic hernia. The exact mechanism of thrombosis was unclear. Amputation of the limb was required and the histopathological findings are presented. This is the first report of such a complication.

Case Report

Arterial thrombosis is an uncommon complication of arterial catheterisation and has not previously been documented in the cat. Here we report a case of ischaemic distal limb necrosis that arose following surgery to correct a peritoneopericardial diaphragmatic hernia.

A 1-year-old, male neutered domestic longhair cat weighing 5.6 kg presented to the Small Animal Hospital, University of Bristol for investigation of acute onset tachypnoea and anorexia.

Physical examination identified elevated temperature (40.1°C), tachycardia [216 beats per min (bpm)], dyspnoea, tachypnoea [60 respirations per min (rpm)] and muffled heart sounds. Mucous membranes were pale pink and dry; capillary refill time was 2 s. Haematology, venous blood gas analysis and serum biochemistry were unremarkable [preoperative packed red blood cell volume (PCV) was 37%]. Conscious thoracic radiographs and ultrasound were consistent with peritoneo-pericardiodiaphragmatic hernia (PPDH).

Initial treatment included intravenous crystalloids (Hartmann’s, 4 ml/kg/h), clavulanic acid potentiated amoxicillin (Augmentin 20 mg/kg IV q6h; GlaxoSmithKline) and oxygen cage therapy for 4 h. The dyspnoea and tachypnoea did not improve with oxygen, therefore surgical intervention was deemed to be necessary.

Intravenous (IV) methadone (0.2 mg/kg, Physeptone; Martindale) and acepromazine (0.01 mg/kg, ACP; Novartis) were administered for analgesia and anxiolysis. Anaesthesia was induced with IV alfaxalone (Alfaxan 0.9 mg/kg to effect; Vetoquinol) and maintained with isoflurane (Isoflo; Abbott) in 100% oxygen via a 5 mm uncuffed endotracheal tube and modified Ayre’s T piece. Monitoring included pulse oximetry, capnography, electrocardiography and intermittent Doppler blood pressure monitoring. Intravenous Hartmann’s solution was infused at 10 ml/kg/h. Manual intermittent positive pressure ventilation was commenced immediately after induction. Ten minutes after induction, the hair over the right metatarsal region was clipped and the skin surgically prepared with chlorhexidine (4%, Hibiscrub; Molnlycke) and alcohol (70%). A 24-gauge (Jelco; Smiths Medical) catheter was successfully placed in the right dorsal pedal artery at the first attempt, secured with microporous tape (3M Micropore), flushed with sterile heparin saline and connected to a transducer for blood pressure monitoring (Beneview T5 monitor; Mindray) via saline filled non-distensible tubing. The transducer was positioned level with the point of the shoulder and once ready for use Doppler measurements ceased.

Midline coeliotomy was performed. The left and right medial liver lobes were identified in the pericardial sac and were drawn into the abdomen. The herniated liver lobes were friable and despite gentle handling, profuse haemorrhage occurred from the parenchyma (blood loss 10 ml/kg estimated from swabs and suction; however, this could be an underestimate owing to pooling in the abdomen). Haemostasis of the liver parenchyma was achieved with oxidised cellulose polymer (Surgicel; Ethicon). The diaphragmatic defect was closed with 3-0 polydiaxonone (PBS: Ethicon) in a simple continuous pattern. During haemorrhage, mean arterial blood pressure (MAP) decreased from 86–96 mmHg to 45 mmHg, and intravenous colloids (6 ml/kg over 20 mins; Gelofusin) were provided. The concentration of isoflurane was reduced and 2 µg/kg of fentanyl was administered intravenously. Despite this, heart rate increased (170 bpm to 215 bpm) and systolic arterial pressure SAP/diastolic arterial pressure (DAP) (MAP) declined to 50/32(38) mmHg. An infusion of bovine polymerised haemoglobin (Oxyglobin; Dechra) was commenced at 7 ml/kg/h and the rate of crystalloids was increased to 80 ml/kg/h. Five minutes later, sinus bradycardia developed (80 bpm), followed by loss of the capnography trace and pulsatile arterial waveform. Pulses were non-detectable manually or with Doppler ultrasound, indicating loss of effective cardiac output. Adrenaline (0.1 mg IV) restored cardiac output and MAP increased, although remaining lower than normal. MAP remained low (45–55 mmHg) for the following 10 mins, despite ongoing fluid therapy. Electrical activity of the heart was not lost; cardiac massage was not required.

By the end of the surgery (45 mins duration) the heart rate (HR) and respiratory rate were within the normal limits (HR: 155 bpm) and SAP/DAP (MAP) was 130/74(90) mmHg. Immediate postoperative PCV was 10% (haematocrit: 10.2%, reference interval 25–45%); total solids were 55 g/l. An abdominal pressure bandage was applied for 8 h because it was not possible to achieve absolute haemostasis from all surfaces of the liver parenchyma. The low postoperative haematocrit may have been explained by IV fluid therapy, intraoperative bleeding and/or continued postoperative haemorrhage. Forty millilitres of whole blood (type-matched A) was provided intravenously postoperatively.

During recovery from general anaesthesia, flushing of the arterial catheter was not possible and it was removed. No bleeding was noted on catheter removal and a compression bandage was applied to the site for 10 mins. No abnormalities (eg, swelling, bruising) were noted.

Medication provided immediately postoperatively included intravenous crystalloids (Hartmann’s, 4 ml/kg/h), methadone (Physeptone 0.2 mg/kg IV q4h for 36 h; Martindale), clavulanic acid potentiated amoxycillin (20 mg/kg IV q6h for 24 h, switching to oral medication for a further 3 days), metronidazole (Metronidazole 10 mg/kg IV q12h for 5 days; Bioflow Baxter), ranitidine (Zantac 2 mg/kg IV q12h for 5 days; GlaxoSmithKline) and vitamin K (Konakion 1 mg/kg SC q12h for 3 days; Roche).

After 24 h, haematology revealed marked neutrophilia (65.51 × 10⁹/l; reference interval 2.4–12.5 × 10⁹/l) with toxic changes, and red blood cell anisocytosis and polychromasia.

Four days postoperatively, the cat demonstrated pyrexia (39.9°C), an acutely painful right hindlimb in the region of the catheter site, with soft tissue swelling from the mid-tibia distally and 8/10 right hindlimb lameness. The dorsal pedal pulse was detectable using Doppler ultrasound, but not appreciable digitally, possibly because of the soft tissue swelling. Radiography of the limb showed no bony abnormalities. Fine needle aspiration (FNA) of the soft tissue swelling was consistent with septic neutrophilic inflammation. A profuse, pure growth of Klebsiella pneumoniae was cultured (sensitive to cefovecin, gentamicin, imipenem, streptomycin and tetracycline). Amoxicillin and metronidazole was discontinued and cefovecin (Convenia 40 mg SC; Pfizer) was provided. Buprenorphine (Vetergesic 0.02 mg/kg IV, q6–8h; Alstoe) and meloxicam (Metacam for cats 0.05 mg/kg PO q24h; Boehringer-Ingelheim) were also administered. Passive range of motion exercises, physiotherapy and warm compresses were provided q4h.

Three days later, no clinical improvement was apparent. The skin over the region of the arterial catheter placement was firm with questionable viability. Repeat FNA yielded K pneumonia, resistant to tetracycline and cefovecin, but sensitive to imipenem and gentamicin. Meropenem (12 mg/kg SC q12h) was administered, but after a further 3 days a 1.5 cm diameter region of skin over the dorsal pedal artery necrosed. The digits were cold, with no deep pain sensation or palpable pulse. Mid-femoral amputation was subsequently performed and the cat recovered uneventfully.

The limb was immersed in 10% neutral buffered formalin and submitted for pathological examination. Upon gross examination, the skin over the dorsal metatarsus and paw was thickened, very firm and partially sloughing. Close to the junction of the non-viable and viable skin of the dorsal metatarsus, and extending from a point approximately 8 mm distal to the divergence of the long digital extensor tendons, the dorsal pedal artery was firm and bluish-brown over a length of approximately 12 mm. Distal to this region, the vessel was poorly defined.

Microscopically, the main histopathological features were of extensive degeneration and necrosis of the dorsal metatarsal skin and subcutaneous tissues accompanied by thrombosis of multiple large and small calibre blood vessels, including large calibre arteries and a vessel corresponding to the anatomical location of the dorsal pedal artery (Figure 1). Other findings included variably mild-to-marked mixed inflammation, granulation tissue formation, and colonisation of necrotic tissue by Gram-negative bacteria and fewer Gram-positive bacteria. These findings were consistent with arterial thrombosis and subsequent ischaemic necrosis of the dependent metatarsal and pedal tissues. No evidence of vascular thrombosis was observed in tissue sections examined from various tissue samples collected from sites proximal to the necrotic skin.

Figure 1

(a) Deep subcutaneous tissues from the proximal dorsal metatarsus just distal to the level of the tarsal–metatarsal joint. The subcutis contains a normal large muscular artery (A) (dorsal pedal artery). Haematoxylin and eosin (HE) stain; bar = 0.5 mm. (b) Deep subcutaneous tissues of the dorsal metatarsus distal to the site shown in (a) and at the junction with the region of gross thickening of the skin. Note the large artery containing a thrombus (T) surrounded by inflamed granulation tissue adjacent to a distal region of necrosis (N). HE stain; bar = 0.5 mm. (c) Skin and subcutaneous tissues from the dorsal metatarsus, distal to the site shown in (b). There is full thickness necrosis of the skin (D) and subcutis (S) accompanied by a marked inflammatory cell infiltrate in the deep subcutis. HE stain; bar = 0.5 mm. (d) Deep subcutaneous tissues from the dorsal metatarsus, deep to the site shown in (c). Degenerate and dilated large muscular blood vessel containing a thrombus (T) and partially surrounded by inflamed granulation tissue. The anatomical location of this blood vessel was consistent with the dorsal pedal artery. HE stain; bar = 0.5 mm

In veterinary medicine, indwelling arterial catheters (IAC) are most commonly placed in the dorsal metatarsal artery, which originates from the cranial tibial artery. Catheters are used for blood pressure monitoring and blood sampling during high-risk surgical procedures and intensive care settings. PPDH corrective surgery carries the risk of complications, including hypotensive episodes, multifocal ventricular premature complexes and loss of palpable pulses; for these reasons, the use of IACs is indicated. ¹

Complications following IAC placement have been reported in horses, ² but, to the best of our knowledge, this case report is the first reported such complication in cats. The major complication rate in humans is less than 1%,^3–5 and necessity for consequential limb amputation is reported. ⁶

Damage to the vessel wall, presence of foreign material and alteration of blood flow may promote thrombus formation after IAC. ⁷ Complications increase with larger catheters, ⁸ placement for longer than 48–72 h or in patients with low cardiac output,^9,10 but not with hypotension. ¹¹ Arterial thrombosis may cause complications, including tissue necrosis, ¹² and in humans is treated with unfractioned heparin, thrombolytic agents (recombinant tissue plasminogen activator), oral anticoagulants (coumarin) and antiplatelet agents (aspirin).

In the case described herein, traumatic placement and prolonged catheterisation did not occur. The catheter was removed less than 2 h after placement and was flushed remotely, making introduction of external contaminants unlikely. Bacterial culture of a cephalic venous catheter (removed at day 4 after the development of pyrexia) did not yield any growth, although this does not exclude bacteraemia. No blood cultures were performed. Arterial thrombosis could have been secondary to haematogenous spread of K pneumoniae liberated from the entrapped liver lobes or gall bladder following surgical manipulation.^13,14 This cannot be confirmed because samples of liver were not submitted for bacterial culture.

It is possible that the loss of peripheral perfusion may have contributed to the development of thrombosis, owing to vascular stasis, hypoxaemia or post-resuscitation disease ¹⁵ in a vessel with pre-existing endothelial trauma. Bacteraemia following cardiac arrest and resuscitation^16–18 may lead to bacterial seeding to thrombi or tissues. This cat suffered a period of ineffectual cardiac output, although cardiac contraction did not stop. For this reason, resuscitation with vasoconstriction (caused by adrenaline) and fluid therapy was successful without requiring cardiac massage. Ventilation of the lungs was provided throughout.

This patient originally presented with an elevated temperature and it is not clear whether this was due to hyperthermia or pyrexia. Despite the initial normal haemogram, it is possible that there was an underlying infectious cause present prior to surgery. No cause was discernible in this case.

With respect to the lameness, which was identified acutely 4 days postoperatively, it is important to note that the cat was rested within a cage for 4 days, during which no abnormalities on palpation of the limb were detected. However, it is possible that had he been allowed to walk around the feline ward before this time, a more subtle lameness would have been detected earlier. If the thrombosis occurred at the time of catheterisation, resulting in subsequent ischaemia and infection, one might expect the lameness to be apparent earlier than 4 days postoperatively. A subtle lameness may have been missed in this case. Alternatively, and supported by the detection of a dorsal pedal flow on Doppler, the thrombosis may have occurred later, secondary to infection in an already traumatised vessel, in which case the lameness would not be detectable before day 4 postoperatively.

Conclusions

Although arterial catheterisation provided useful and continuous information during anaesthesia to guide resuscitation, the procedure is not benign and may have contributed to later complications in the limb. This is the first reported such complication in cats. Arterial thrombi may be an additional consideration where surgical manipulation of the liver is performed. In the case described, clinical deterioration occurred in the face of appropriate culture-guided antibiotic therapy and amputation was required.