Management of acute kidney injury with continuous venovenous haemodiafiltration in a cat

Abstract

Continuous renal replacement therapy is an emerging technique for the treatment of acute kidney injury (AKI). Data regarding its use in cats are limited. This report describes the use of a novel continuous renal replacement therapy (CRRT) system for the treatment of AKI in a cat. A 1.3-year-old cat developed uraemic signs following the administration of a non-steroidal anti-inflammatory agent for the treatment of a suspect traumatic episode. CRRT was provided with a Prismaflex Gambro machine used in continuous venovenous haemodiafiltration mode, with an AN-69 surface-treated membrane, synthetic colloid priming and heparin anticoagulation. Two treatment cycles were performed, totalling 51 h of CRRT. The treatment was effective in controlling uraemic signs, and no major complications were noted. Owing to financial constraints the owners declined further CRRT treatments, and on day 8 of hospitalisation, owing to the lack of significant clinical improvement, humane euthanasia was performed. The set-up detailed in this report provides a viable option for the initial treatment of cats with AKI.

Case Report

Acute kidney injury (AKI) refers to a rapid decrease in renal function, usually within 48 h. ¹ AKI is a commonly encountered emergency presentation in veterinary patients and is associated with a poor prognosis.^2–4 Renal replacement therapy (RRT) is indicated when medical management fails and AKI results in intractable uraemic signs, and life-threatening acid–base, electrolyte and fluid disturbances. ⁵ Continuous RRT (CRRT) describes a spectrum of blood purification techniques that have been recently reported for the treatment of AKI in both dogs and cats.^6–8 While intermittent haemodialysis (IHD) is mainly a diffusive modality, CRRT can combine convection and diffusion to achieve solute clearance, allowing for removal of larger molecules. ⁹ Moreover, CRRT allows a gradual correction of electrolyte, acid–base and body fluid imbalances, promoting the maintenance of haemodynamic stability. Despite these theoretical advantages, current human literature failed to identify a clear superiority in outcome when comparing the two techniques. ¹⁰ In feline patients, CRRT has been associated with a number of complications, and experience with this technique is limited.^6,7 This report describes the use of venovenous haemodiafiltration in a feline patient with AKI.

A 1.3-year-old, 5 kg, male neutered domestic shorthair cat presented to the referring veterinarian following a possible traumatic episode. Earlier in the day the owners found the cat to be very lethargic and noted the presence of several superficial excoriations. The cat had access outdoors, but no other previous medical history or access to toxins was reported. Physical examination revealed no abnormalities other than the superficial lesions. The cat was treated with a long-acting antimicrobial, cefovecin 10 mg/kg (Convenia; Zoetis) and a single dose of a non-steroidal anti-inflammatory drug, meloxicam 0.2 mg/kg (Metacam; Boehringer Ingelheim), both administered subcutaneously (SC). In the days following discharge the cat began vomiting and, 5 days from initial presentation, developed ataxia. The cat represented to the referring veterinarian where physical examination revealed obtundation, bradycardia, renomegaly and abdominal pain. Serum biochemistry showed marked hyperkalaemia (9.4 mmol/l; reference interval [RI] 3.5–5.8 mmol/l) and azotaemia (urea >46.4 mmol/l; RI 5.7–12.9 mmol/l). A diagnosis of AKI was made and treatment was initiated with high-rate fluid therapy and diuretics (2 mg/kg furosemide IV). Following 6 h of treatment no measurable urinary output was observed and the cat was referred to the Queen Mother Hospital for Animals Emergency Service for further investigation and management.

On presentation, the patient was obtunded and hypothermic (36.6°C). Peripheral pulses were hypodynamic but synchronous. Heart rate and respiratory rate were 160 beats per min (bpm) and 24 breaths per mins, respectively. Body weight was 5.3 kg, with a body condition score of 5/9. Abdominal palpation confirmed bilateral renomegaly and marked abdominal pain. Mild chemosis was also present bilaterally. Non-invasive blood pressure (NIBP) measured by Doppler technique was 180 mmHg. Continuous electrocardiography showed signs consistent with hyperkalaemia: bradycardia with an average heart rate between 130 and 160 bpm, atrial standstill, widened and depressed QRS complexes and increased amplitude of T waves. Echocardiography revealed no evidence of left ventricular hypertrophy, but right and left atria were enlarged, likely secondary to fluid overload. Moderate volume pleural and small volume pericardial effusions were also observed. Abdominal ultrasonography demonstrated mild bilateral renal enlargement (5 cm in length). The renal cortices and medullae were hyperechoic and there was decreased corticomedullary differentiation. The renal outlines were slightly irregular. Mild bilateral pyelectasia was present. A small volume of anechoic retroperitoneal and peritoneal fluid was also noted. The urinary bladder was small, but there was urine present. A urine sample was collected by cystocentesis for urinalysis and urine bacterial culture. Urinalysis revealed a urinary specific gravity of 1.015, a urinary pH of 6, trace of proteinuria, trace of glucosuria and 4+ haematuria. Urine sediment analysis revealed 10–20 red blood cells per high-power field (HPF), one to five white blood cells per HPF, occasional epithelial cells and some organic debris. Serum biochemistry abnormalities included severe hyperkalaemia, ionised hypocalcaemia, hypermagnesaemia, hyperglycaemia, hyperphosphataemia and azotaemia (Table 1). Venous blood gas showed marked metabolic acidosis (Table 1).

Table 1

Serial evaluation of selected laboratory and monitoring data

	Reference interval	Admit	Start first CRRT	End first CRRT		Start second CRRT	End second CRRT
		Day 0	Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7	Day 8
pH	7.36–7.47	7.12	7.18	7.40	7.37	7.43	7.43	7.46	7.42	7.45
HCO3- (mmol/l)	14.0–22.0	9.7	14.8	23.0	21.7	19.1	25.9	19.8	18.5	17.4
Base excess	–4.0 to +4.0	–19.7	–13.8	–2.0	–3.7	–5.3	1.4	–4.3	–6.2	–6.8
Sodium (mmol/l)	140.0–153.0	145.0	146.7	140.6	139.2	141.5	143.6	146.2	146.8	145.5
Chloride (mmol/l)	106.0–120.0				110.5	113.5	112.3	115.3	116.6	118.1
Potassium (mmol/l)	3.60–4.60(3.70–5.80)	8.49 (11.00)	9.00 (8.60)	4.76 (5.80)	5.20	4.65 (5.00)	5.43 (6.20)	4.81	4.71	4.57
Ionised calcium (mmol/l)	1.13–1.33	1.05	1.07	1.16	1.11	1.12	1.17	1.19	1.17	1.14
Ionised magnesium (mmol/l)	0.35–0.55	0.63	0.58	0.47	0.46	0.55	0.40	0.42	0.49	0.46
Glucose (mmol/l)	4.20–6.60	7.80	9.26	5.18	4.11	4.26	4.71	4.60	3.98	3.67
Urea (mmol/l)	3.0–10.0(3.6–10.7)	(103.8)	(97.0)	19.3 (17.8)		(43.5)	10.8 (11.4)	19.8	27.6	30.7
Creatinine (µmol/l)	50–140(27–186)	(2576)	(2048)	520 (380)	845	1191 (1007)	346 (299)	842	923	926
Phosphorus (mmol/l)	(1.10–2.74)	(8.25)	(6.29)	(1.59)		(3.19)	(1.30)
PCV (%)	24–45	42	35	25	29	23	20	22	20	20
Total plasma protein (g/l)	60–80	75	60	60	64	63	69	62	60	60
Body weight (kg)		5.31	5.39	5.15	5.32	5.30	4.84	4.89	4.94	5.01
NIBP (mmHg)		180	140	180	180	153	120	160	140	130
UOP (ml/kg/h)			0.27	0.40	0.23	0.43	0.39	0.33	0.30	<0.50

Values enclosed in parentheses were obtained with an in-house serum biochemistry analyser (Vetscan; Abaxis). Packed cell volume (PCV) was manually measured. Total plasma protein was measured with a hand-held refractometer (Burtons). All other laboratory data were measured with an in-house blood gas analyser (phOx Ultra; NOVA Laboratories)

CRRT = continuous renal replacement therapy; NIBP = non-invasive blood pressure, measured by Doppler technique at the metacarpal artery (Ultrasonic Doppler 811B Flow Detector; Parks Medical Electronics); UOP = urinary output production

Hyperkalaemia was treated with the administration of calcium gluconate (calcium gluconate injection 10%; Hameln Pharmaceuticals) 0.5 ml/kg intravenously (IV), regular insulin (Actrapid; Novo Nordisk) 0.5 IU IV, dextrose (glucose 25 g/50 ml; Hameln Pharmaceuticals) 0.5 g/kg IV, bicarbonate (8.4% Sodium Bicarbonate; B. Braun) 1 mEq/kg IV and terbutaline (Brecanyl; AstraZeneca) 0.01 mg/kg SC. These therapies were repeated as needed in the first 24 h of hospitalisation to control hyperkalaemia. The cat also received 0.9% saline (Vetivex 1; Dechra) supplemented with 2.5% dextrose at 2 ml/kg/h as initial IV fluid therapy.

An indwelling urinary catheter was placed and connected to a closed collection system for urinary output monitoring. Given the high clinical suspicion of volume overload and oliguria or anuria, a furosemide 2 mg/kg IV bolus (Dimazon; Intervet) was administered, followed by a continuous rate infusion (CRI) at 0.5 mg/kg/h. Owing to concerns of a possible urinary tract infection, antimicrobial therapy with potentiated amoxicillin (Augmentin; GlaxoSmithKline) 20 mg/kg IV every 8 h, was initiated while pending the results of the urine bacterial culture. Analgesia was initially provided with methadone (Physeptone; Martindale Pharmaceuticals) 0.1 mg/kg IV every 4–6 h and then a fentanyl (Sublimaze; Janssen-Cilag) CRI at 0.03–0.06 μg/kg/h titrated based on pain assessment. Omeprazole (Losec; AstraZeneca) 0.5 mg/kg IV once daily and maropitant (Cerenia; Zoetis) 1 mg/kg SC once daily were also administered. The cat was documented to be oliguric, with a urinary output in the first 24 h of hospitalisation of 0.27 ml/kg/h. Owing to persistent overhydration, oliguria, hyperkalaemia and acid–base disturbances unresponsive to medical management, we elected to perform CRRT. Two CRRT cycles were performed at 24 and 96 h after admission, 48 h apart. The two treatments were continued for 24 and 27 h, respectively. Treatments were discontinued once acid–base imbalances and hyperkalaemia resolved, and adequate fluid removal and urea reduction were achieved. A 48 h interval between treatments was selected based on clinical assessment of the patient and sequential evaluation of laboratory data.

To provide vascular access, an 8 Fr 10 cm short-term dual lumen haemodialysis catheter (Kflow Epic; Kimal) was percutaneously placed in the right jugular vein under sedation with ketamine 5 mg/kg IV (Ketaset; Fort Dodge) and midazolam 0.25 mg/kg IV (Hypnovel; Roche). Correct positioning of the tip of the catheter at the junction between the right atrium and cranial vena cava was confirmed by fluoroscopy. During the first CRRT cycle the configuration of the vascular access was changed to prevent frequent treatment interruption: the return port on the dialysis catheter was used to remove blood from the patient and a medial saphenous catheter (MILACATH Guidewire – Single Lumen 6 Fr 15 cm; MILA), initially placed as sampling line, was used to return blood to the patient. This configuration was also used for the second treatment.

Venovenous haemodiafiltration was initiated. This modality utilises a combination of convection (haemodialysis) and diffusion (haemofiltration) to achieve the clearance of solutes. CRRT was delivered with a new generation extracorporeal blood purification system (Prismaflex; Gambro) using a haemodiafilter (ST 60; Gambro) containing a surface-treated acrylonitrile and sodium methallyl sulfonate copolymer (AN69) membrane (Figure 1). The surface treatment of the membrane allows it to absorb the heparin present in the priming solution thus reducing the risk of filter clotting. According to manufacturer’s specification, the filter nominal effective surface area is 0.6 m², the filter volume is 44 ml ± 10% and the total extracorporeal volume within the set is 93 ml ± 10%. Before initiation of the treatment the set was primed with a solution of 0.9% saline (Vetivex 1; Dechra) and unfractionated heparin (UFH) at 5000 IU/l (heparin sodium 1000 IU/ml; Wockhardt). The extracorporeal volume represented approximately 30% of the calculated circulating blood volume of the treated patient. To prevent problems associated with this large extracorporeal volume and to remove any circulating heparin not bound to the filter, additional priming was performed with a hydroxyethyl starch solution (Voluven; Fresenius-Kabi).

Figure 1

Haemodiafilter (ST60; Gambro) used during treatment of a cat with acute kidney injury. The surface treatment of the AN69 membrane with polyethyleneimine neutralises anionic surface charges. This improves biocompatibility by reducing contact activation of the intrinsic coagulation pathway and by promoting the binding of heparin in the priming solution to the membrane surface

Blood flow rate was maintained at 20 ml/min (4 ml/kg/h). Dialysate flow rates ranged between 50 and 250 ml/h (10–50 ml/kg/h). The ultrafiltration flow rate used was 150 ml/h (30 ml/kg/h). Replacement fluids were infused two-thirds as pre-dilution (before the filter) and one-third as post-dilution (after the filter). Fluid removal rates of 15–50 ml/h (3–10 ml/kg/h) were used, with a progressive increase in the fluid removal rate throughout the procedure. A commercial bicarbonate-buffered solution (Prismasol 4; Gambro) was used both as dialysate and replacement fluid. This solution has a sodium concentration of 140 mmol/l. To prevent hyponatraemia and to reduce the risks associated with a too rapid a reduction in plasma osmolarity, both replacement and dialysate solutions were spiked with hypertonic saline 7.5% (VetCare hypertonic NaCl; B. Braun) to achieve a predicted sodium concentration of 145 mmol/l.

Anticoagulation was achieved with systemic administration of UFH and monitored with an activated clotting time (ACT) point-of-care analyser (Hemochron Signature+; ITC). An initial IV bolus of 50 IU/kg UFH was followed by a CRI started at 10 IU/kg/h and then titrated according to previously published guidelines to achieve a target ACT of 150–180 s. ⁹ ACT was measured every 15 mins until achievement of a steady state and then every 2–4 h. UFH CRI dose range was 10–33 IU/kg/h. ACT values ranged between 140 s and 454 s. However, the majority of measured ACT values were within 150 and 200 s. No evidence of bleeding was observed, and visual inspection of the filters following the treatments did not reveal any signs of clotting.

Temperature was continuously monitored and maintained using a forced-air warming system (Bair Hugger; Arizant) prior to the initiation of CRRT and continued throughout treatment to maintain normothermia. A blood-warmer device was also used to warm blood in the return bloodline (Prismacomfort; Gambro). Temperature remained within normal limits during both treatments. Oxygen supplementation via flow-by was provided until the patient fully recovery from the sedation administered to facilitate catheter placement.

NIBP was monitored hourly during CRRT by Doppler technique at the metacarpal artery (Ultrasonic Doppler 811B Flow Detector; Parks Medical Electronics). NIBP ranged between 110 mmHg and 190 mmHg. The initiation of CRRT was associated with a drop in NIBP of 20 mmHg and 30 mmHg on the first and second treatment, respectively. This decrease in blood pressure was transient and did not require treatment. No episodes of hypotension were recorded during or in between treatments. Hypertension (NIBP >180 mmHg) developed between treatments and was treated with oral amlodipine (Istin; Pfizer) at 0.625 mg every 24 h.

A dry body weight of 4.8–5.0 kg was estimated based on the medical record from the referring veterinary practice. With a body weight on presentation of 5.30 kg, the patient was assessed as being 6–10% overhydrated. Body weight decreased from 5.39 kg to 5.15 kg during the first CRRT cycle and from 5.30 kg to 4.84 kg during the second CRRT cycle (Table 1). Repeated bedside ultrasonography following the first CRRT cycle showed complete resolution of the tricavitary effusion noted on initial presentation. Fluid therapy was titrated after initial presentation based on fluid balance, body weight and hydration status. Following the first CRRT cycle, IV fluids were switched to compounded sodium lactate, a balanced isotonic crystalloid solution (Vetivex 11; Dechra). The furosemide CRI was discontinued during CRRT.

Acid–base status, packed cell volume (PCV), total plasma protein and serum biochemistry parameters were regularly monitored during and between treatments with point-of-care analysers (PhOx Ultra; NOVA Biomedicals – Vetscan; Abaxis). Relevant results are summarised in Table 1. To prevent the development of hypophosphataemia, a phosphate (Addiphos; Fresenius-Kabi) CRI was started at 0.012 mmol/kg/h during the second CRRT cycle.

The cat appeared to tolerate the treatment well and no evidence of neurological or respiratory complications was observed. Owing to concerns that rapid administration of the large extracorporeal volume could lead to volume overload, the blood remaining in the set at the end of the treatments was transferred to a transfusion bag prefilled with citrate (ACD-A; Haemonetics) with a 1:9 citrate to blood ratio and immediately aseptically transferred into sterile 20 ml syringes, which were stored at 4°C until the time of administration. Sequential administration of stored blood in syringes was then performed over the following 6 h.

Despite the lack of any obvious indication of blood loss or filter clotting, the cat developed anaemia (Table 1). No evidence of haemolysis was seen and the anaemia appeared to be non-regenerative on blood smear evaluation. We speculate that the anaemia was likely multifactorial in origin, due to a combination of blood loss, decreased red blood cell survival and decreased erythropoiesis caused by renal disease, inflammation and uraemia. Blood loss might have occurred through occult gastrointestinal bleeding, thrombocytopathia, or incomplete recovery of the extracorporeal blood at the end of the treatments. Frequent blood sampling might have also contributed to the anaemia, with approximately 5% of the circulating blood volume removed for diagnostic testing. No clinical signs associated with the anaemia were observed; however, given the magnitude of the reduction in PCV, a blood transfusion would have been required if we had proceeded with a further CRRT treatment.

Two common methods used to evaluate renal replacement treatment dose and adequacy are Kt/V and urea reduction ratio (URR). Kt/V is a dimensionless number that considers urea clearance, duration of treatment and urea volume of distribution. A Kt/V >1.4 has been shown to be associated with improvement in outcome in humans with AKI. ⁹ The settings used in our cat resulted in a calculated Kt/V ranging between 1.5 and 3.2. The URR is the percentage of reduction of the urea during the treatment. Overall URR was 0.82 and 0.74 during the first and second treatment, respectively. Hourly URR never exceeded 0.05/h. An hourly URR <0.1/h has been recommended in IHD treatments to prevent complication such as dialysis disequilibrium syndrome. ¹¹

The urine bacterial culture submitted on presentation revealed no growth of aerobic or anaerobic bacteria. The owners declined renal biopsies or fine-needle aspirates. Urinary output remained <0.5 ml/kg/h throughout hospitalisation. Owing to financial constraints the owners declined further CRRT treatments, and on day 8 of hospitalisation, owing to the lack of significant clinical improvement, humane euthanasia was performed. Permission for post-mortem examination was not given. The aetiology of the AKI remained undetermined in our case. However, given the history and the time of onset of clinical signs, non-steroidal anti-inflammatory drug-related toxicity represents a likely differential diagnosis. Given the changes in renal appearance on abdominal ultrasonography, acute-on-chronic kidney disease cannot be ruled out.

CRRT has been previously described as a viable alternative to IHD for the management of AKI in cats.^6,7 These previous reports describe the use of an older CRRT machine, which has since been discontinued (Prisma; Gambro), along with citrate regional anticoagulation. This older machine has been replaced with the model used in this case report (Prismaflex; Gambro). Compared with previous CRRT systems, the Prismaflex allows greater flexibility, with the possible use of simultaneous haemodialysis and haemofiltration and fluid replacement both in pre- and post-dilution configurations. Despite the large extracorporeal volume required, no evidence of hypotension was noted with the use of a colloid solution for priming. The use of a synthetic colloid solution was deemed necessary owing to the lack of readily-available stored feline blood products at our institution. The colloid used in the current report has been recently associated with a worse outcome and a higher incidence of AKI in critically ill humans,^12,13 and for this reason is currently not available in the UK. Alternatives, such as haemoglobin-based oxygen carriers (Oxyglobin; Dechra) or gelatins (eg, Gelofusine; B Braun), may have to be considered in the future. The use of heparin anticoagulation and a surface-treated membrane allowed for adequate anticoagulation with no signs of clotting in the circuit.

To our knowledge, this is the first report of the use of a novel blood purification system (Prismaflex; Gambro) along with a surface-treated haemodiafilter, heparin anticoagulation and colloid priming for the treatment of AKI in a feline patient. The set up used in our case allowed for an effective and safe treatment with minimal complications.

Approximately 50% of cats treated for AKI with renal replacement therapies survive to discharge, with survival rates similar to those reported in humans.^7,14,15,16 The outcome appears to be related to the underlying aetiology of the AKI, and toxicities are associated with a worse prognosis. ¹⁷ The number of treatments required to achieve the intended effect varies greatly: two different studies evaluating IHD for the treatment of AKI in cats report three as the median number of treatments performed, ranging from one to 34.^14,15

Conclusions

It is clear from these data that renal replacement therapies require a significant emotional and financial commitment from the owners, but remain currently the only therapeutic option for patients with AKI that fail to respond to medical management. Current veterinary literature regarding renal replacement therapies mainly describes IHD techniques. Further studies are needed to establish the role of CRRT in the treatment of veterinary patients with AKI and develop guidelines for its clinical application.