This paper was presented as an abstract at the ACVIM Forum 2008.
Corresponding author: Mary Anna Labato, Foster Hospital for Small Animals, Cummings School of Veterinary Medicine, Tufts University, 200 Westboro Road, North Grafton, MA 02135; e-mail: firstname.lastname@example.org.
Background: Peritoneal dialysis (PD) has been described for use in animals with acute kidney injury refractory to fluid therapy. However, no study has examined the use of PD in a large group of cats.
Hypothesis: PD is an important adjunctive therapy to treat acute kidney injury in cats.
Animals: The medical records of 22 cats with acute kidney injury that had received PD were examined. Animals were excluded if acute uremia was a result of postrenal causes such as uroabdomen or urethral obstruction.
Methods: Medical records were reviewed for the following: indication for PD, outcome, number of cycles performed, survival time, and predialysis and postdialysis results for blood urea nitrogen (BUN), creatinine, potassium, chloride, sodium, phosphorus, total protein, and albumin concentrations, and urine output.
Indications for PD include acute-on-chronic kidney injury, acute kidney injury caused by toxins, bilateral ureteroliths, bilateral ureteral ligation as a complication of ovariohysterectomy, and unknown causes. The median survival time for all cats on PD was 4 days, although the median survival time for the cats that were discharged was 774 days. The most common complications were dialysate retention and sequestration of dialysate SC. There was a significant (P< .05) decrease between predialysis and postdialysis results for BUN, creatinine, potassium, phosphorus, total protein, and albumin concentrations. There was a significant (P< .05) difference in survival times between sexes.
Conclusions and Clinical Importance: PD is an effective option for treatment of cats with acute kidney injury refractory to fluid therapy.
Peritoneal dialysis (PD) is the process of utilizing the peritoneum as a semipermeable membrane in order to move solutes and water between blood in the peritoneal capillaries and fluid (dialysate) instilled into the peritoneal cavity.1 PD most frequently is used in management of acute kidney injury refractory to fluid therapy, but it also has been used in management of severe metabolic disturbances, acute poisoning with dialyzable substances (eg, ethylene glycol, ethanol, barbiturates), and severe temperature extremes.2
PD is performed by placing a catheter into the abdomen that is attached to a closed collection system. A specialized solution (dialysate) is infused into the abdomen, allowed to dwell for a predetermined amount of time, and then removed. The dialysate can be formulated in the clinic, or can be purchased commercially. Standard commercial dialysate solutions are designed to remove urea, creatinine, potassium, and phosphate from the plasma into the dialysate by the process of diffusion. A variety of dialysate solutions can be used with differing osmolalities on a case-by-case basis depending on the fluid balance of the patient. Fluid and solutes move across the peritoneum by diffusion, ultrafiltration, and convection. Urea and potassium diffuse across the peritoneal membrane quickly, whereas creatinine and phosphorus take longer to equilibrate.3 In humans receiving hemodialysis and PD, urea kinetics are evaluated to assess the adequacy of treatments.
Several studies have reported on the technique of PD in dogs.4–6 One study examined PD in both dogs and cats, but only 2 cats were included in the study, and neither cat survived.5 Another study recently evaluated PD in 6 cats.7 That study showed a better outcome in cats receiving PD than did the previous study.5,7 The smaller population size in this study made it difficult to evaluate trends and prognostic indicators for survival. In addition, the use of negative pressure during the outflow phase of each PD cycle, which has not been described previously in the veterinary literature, may have resulted in different complications and outcomes of these patients compared with those treated by standard methods of PD.7 The present study examined the indications, effectiveness, outcomes, and complications associated with the use of PD in cats with acute kidney injury. It also evaluated the urea reduction ratio (URR) as a measure of dialysis adequacy and predictor of survival in cats receiving PD.
Materials and Methods
Criteria for Selection of Cases
Medical records of 27 cats that had received PD at Cummings School of Veterinary Medicine at Tufts University from January 1, 2001 to December 31, 2006 were reviewed. Criteria for inclusion in the study consisted of diagnosis of acute kidney injury and at least 1 PD cycle performed. Acute kidney injury was defined as a severe and sudden decrease in glomerular filtration rate and subsequent uremia. This study included cats with acute kidney injury caused by toxin exposure, obstruction secondary to ureteroliths or ureteral ligation, acute exacerbation of chronic kidney injury, or unknown causes. Five cats with uroabdomen because of bladder rupture were excluded from the study. Twenty-two cats met the criteria for inclusion into the study.
The following data were collected from each patient record and recorded in a systematic fashion: sex, breed, age, weight, indication for PD, number of dialysis cycles, number of days PD was performed, outcome (discharge, euthanasia, or death), days from the start of PD to endpoint (discharge from hospital versus euthanasia or death), survival time (days) from onset of PD, and method of catheter placement (surgical placement with omentectomy versus nonsurgical placement). Complications noted from each patient's record (obtained from daily physical examination, treatment sheet, and dialysis flow sheet) also were recorded.
Predialysis and postdialysis results were obtained for the following variables: blood urea nitrogen (BUN), creatinine, phosphorus, sodium, potassium, chloride, total protein, albumin, and urine output (mL/kg/h). When available, blood glucose concentrations were recorded for the duration of dialysis. In those cases in which cats died or were euthanized before completion of dialysis, the postdialysis results were those obtained closest to death or euthanasia. In 1 cat that experienced cardiac arrest after the 1st PD cycle, postdialysis results were not obtained.
When possible, URR was calculated for each of the cats at 6, 12, 24, 36, 48, and 72 hours after starting dialysis by the following formula:
where the BUNpre is the predialysis BUN and BUNpost is the BUN recorded at specific time intervals.
Mean, median, and standard deviation were calculated for each variable noted above. With the exception of urine output predialysis and postdialysis, all results were compared by a paired t-test. Because urine output data were not evenly distributed a signed-rank test was used. Kaplan-Meier estimates of survival times from onset of PD for age, sex, indications for PD, predialysis results for BUN, creatinine, potassium, phosphorus, sodium, chloride, total protein, albumin, and urine output were computed by a commercial statistical software package.a Mean and median survival time was calculated by the area under the curve for each variable. The Mantel-Cox log-rank analysis was performed to compare the survival distributions for different levels of the variables above. The following divisions were made in comparing different variables: age: <8, ≥8; number of cycles: ≤10, 10–50, ≥50; BUN: <150, 150–250, >250 mg/dL; creatinine: <8, 8–16, >16 mg/dL; phosphorus: <10, 10–20, >20 mg/dL; sodium: <140, 140–149, >149 mEq/L; chloride: <119, ≥119 mEq/L; potassium: ≤8, >8 mEq/L, total protein: ≤6, >6 g/dL; albumin: ≤2.8, >2.8 g/dL; urine output: <1, ≥1 mL/kg/h. Mann-Whitney rank-sum tests were used to compare URR between survivors and nonsurvivors at different time periods.
Of the 22 cats examined, 14 were castrated males and 8 were spayed females. A variety of breeds were represented, including Domestic Short Hair (n = 14), Domestic Long Hair (n = 3), Siamese (n = 2), Russian Blue (n = 1), Himalayan (n = 1), and Maine Coon (n = 1). The average age at presentation was 7.2 years (range, 9 months–17.4 years).
The most common indication for PD was acute-on-chronic kidney injury (7/22; 32%). Urolithiasis accounted for 23% (5/22) of cases, 18% (4/22) of cats presented with acute kidney injury because of toxicity, 3 of which were a result of lily toxicity, and 1 was an unknown toxin. Acute kidney injury was attributed to spay complications in 14% (3/22) of cats (bilateral ureteral ligation in 2 cats and unknown in 1 cat). Fourteen percent (3/22) of the cats presented with acute kidney injury of unknown cause.
Response to PD
There was a significant difference (P< .05) in mean predialysis and postdialysis results for BUN (P< .001), creatinine (P= .003), phosphorus (P= .03), potassium (P= .03), total protein (P< .001), and albumin (P< .001) concentrations (Table 1). There was no significant difference in predialysis and postdialysis results for sodium (P= .55) and chloride (P= .26) concentrations, and urine output (P= .21).
Table 1. Summary of serum creatinine, blood urea nitrogen (BUN), total protein, and albumin concentrations, and urine output in 22 cats treated with PD.
Mean Values at Discharge
PD, peritoneal dialysis.
Total protein (g/dL)
Urine output (ml/kg/h)
The average number of PD cycles performed was 33.7 (median, 17; range, 1–238). Of 22 cats, 10 cats (45.5%) were discharged from the hospital after PD. Seven cats (31.8%) were euthanized before discharge from the hospital and 5 cats (22.7%) died in the hospital while receiving PD. One cat was discharged from the hospital with the PD catheter in place, and came back daily for PD catheter irrigation for 10 days in the event that PD would be necessary again. The PD catheter in this cat was removed 10 days after discharge from the hospital after laboratory test results showed that azotemia was being controlled by SC fluid administration.
Predictors of Survival
By a Kaplan-Meier curve and the Mantel log-rank test, there was a significant difference (P= .038) in survival time between males and females. Five of 8 females (62.5%) survived to discharge, whereas only 5/14 (35.7%) males survived to discharge.
Indication for PD
There was no significant difference (P= .09) in survival time regardless of the indication for PD by the Mantel log-rank test. All 3 cats with spay complications were discharged from the hospital. None of the cats with spay complications had urine cultures performed within 48 hours of admission to the hospital.
Sixty percent (3/5) of cats with urolithiasis-induced acute kidney injury were discharged from the hospital. Of the 2 cats in this group that did not survive, 1 was euthanized and 1 died. Eighty percent (4/5) of cats with urolithiasis had urine cultures performed within 48 hours of admission. None of these urine cultures were positive for bacterial growth.
Three of 7 (42.8%) cats with acute-on-chronic kidney injury were discharged from the hospital. Three of the 4 cats that did not survive to discharge were euthanized and 1 died. Urine cultures were performed within 48 hours of presentation to the hospital in 42.8% (3/7) of cats with acute-on-chronic kidney; none of the urine cultures were positive for bacterial growth. One of the cats that had a negative urine culture on admission had a history of a positive urine culture (Escherichia coli) 2 weeks before presentation, and had been treated with antibiotics. One additional cat had a urine culture performed by a referring veterinarian the day before presentation, but those culture results were not available. End-stage kidneys were noted on necropsy in the 2 cats that did not have urine cultures performed on admission.
Of the cats presented for toxicities, there were no survivors (0%, 0/4). Three of the toxin exposures were confirmed to be lily toxicity; 1 cat had evidence on necropsy of nephrotoxin exposure. Three of these cats were euthanized and 1 died (Fig 1). All cats presented with toxicities had urine cultures performed within 48 hours of presentation and 50% had positive urine cultures. Both of these cultures were positive for E. coli.
Predialysis results for BUN, creatinine, potassium, phosphorus, total protein, albumin, and urine output were not significant predictors of survival time by Mantel log-rank test (P > .05).
Urea kinetics could be determined for a small number of animals at each time point. There was no significant difference in URR between survivors and nonsurvivors at 12 hours (P= .083), 48 hours (P= .99), or 72 hours (P= .564). There was a significant difference (P= .032) between survivors and nonsurvivors for URR at 24 hours postdialysis. The median URR for survivors at 24 hours was 44.2 compared with 6.6 for nonsurvivors (see Table 2).
Table 2. Summary of URR in surviving and nonsurviving cats with acute kidney injury treated with PD.
The most common complication of PD was dialysate retention, which affected 77% (17/22) of the cats. Fifty percent (11/22) of the cats had sequestration of dialysate under the skin, 13.6% (3/22) had obstructed PD catheters, 9% (2/22) had leakage of fluid from the catheter site, and 9% (2/22) had abdominal pain. One cat developed septic peritonitis caused by Klebsiella pneumoniae (4.55%), 1 cat developed hemorrhagic dialysate and subsequently experienced cardiac arrest, 1 cat had dehiscence of its surgical incision, and 1 cat developed pleural effusion.
Hypoproteinemia was the most common laboratory abnormality noted, with 90% (18/20) of cats exhibiting low total protein concentration after undergoing PD. Hypoproteinemia was present in 38.1% (8/20) of the cats before starting PD. Sixteen percent (3/18) of the cats were hypoalbuminemic before starting PD; posttreatment, this number increased to 68.8% (11/16). Hyponatremia was seen in 13/22 cats (59.1%) before PD and in 13/21 (61.9%) after PD. Hypochloremia was identified in 77.3% (17.22) of the cats before PD and in 85.7% (18/21) after PD. Hypokalemia was not noted in the post-PD laboratory evaluation in the present study.5 Hyperglycemia was identified in 3/22 (13.6%) of the cats during the course of PD.
Placement of PD Catheters
Twelve cats (54.5%) had surgical placement of their PD catheters with partial omentectomy. Of these cats, 7/12 (58.3%) had retained dialysate or an obstructed catheter as a complication noted in the medical record. Half of the animals with surgical placement of the catheter (6/12) had SC leakage of dialysate as a complication.
Eight cats (36.4%) had percutaneous placement of PD catheters. One hundred percent of these animals had retained dialysate, an obstructed PD catheter, or both noted as a complication and 5/8 (62.5%) had SC leakage of dialysate as a complication.
Two cats (9.0%) initially had percutaneous placement of their PD catheters, but eventually required surgical placement of PD catheters caused by catheter outflow complications. One of these cats had a surgically placed catheter, but no partial omentectomy and continued to have complications involving retained dialysate. This cat was reoperated and a partial omentectomy performed and fewer complications were reported postoperatively.
Not all records noted the type of catheter placed. Several different types were noted, including Quinton Pediatric Peritoneal Dialysis catheter,b Malecot catheter,c Swan Neck Missouri Peritoneal Dialysis catheter,d and the T-style fluted catheter.e The percutaneously placed types of PD catheters had a much higher occurrence of dialysate retention.
Although overall survival time was poor, animals that are considered for PD are severely ill and have not shown signs of improvement with conventional therapy. Also, the majority of cats that did not survive to discharge were euthanized, and time and financial commitment may have factored into the decision to euthanize.
There was a significantly longer survival time for females than males in this study. One factor that may contribute to this observation is the group of 3 cats that had bilateral ureteral ligation after ovariohysterectomy. All 3 of these cats had surgical correction of this complication in addition to PD. These cats experienced acute kidney injury at a younger age than did the animals with other indications for PD. Therefore it is not surprising that these cats had longer survival times than did the other groups.
In this study, all 4 cats with toxicity did not survive. This is a lower survival rate than in previous studies on Easter lily toxicosis in cats.8,9 The lower rate of survival may have been because of increased severity of the cases that were considered for PD. The damage sustained by the kidneys at such a late stage may be irreversible by the time PD is initiated. A previous study reported the use of hemodialysis in management of 2 cases of acute kidney injury caused by lily ingestion in cats.8 Both of the cats with disease serious enough to warrant hemodialysis died in the hospital. The poor prognosis and expense of care also may have influenced the decision for euthanasia in these cases as well. The small number of cats in this study necessitates cautious interpretation of these results.
The complication rate was relatively high, although most complications were not life threatening. Dialysate retention, sequestration of dialysate under the skin, and obstructed PD catheters were the most common complications.
Peritonitis is a well-described complication of PD.3,10,11 A previous study showed an increased risk of peritonitis with increasing number of days on PD, which was true for the present study as well.5 The 1 cat that was diagnosed with peritonitis was the patient that spent the longest time on dialysis (10 days) and had the largest number of cycles (238). The bacteria cultured in this case of peritonitis was Klebsiella, which was found to be one of the most common bacterial isolates in a previous study.5
Pleural effusion was a complication seen in 1 cat in this study. Pleural effusion has been described previously as a complication of PD in a dog, and was suggested to be because of a pleuroperitoneal connection.12 Glucose concentration of the pleural effusion was not measured in the current case, so it was not possible to differentiate whether the pleural effusion was because of a pleuroperitoneal connection or fluid overload.
Although the catheter outflow complication rate was high for both percutaneous and surgically placed catheters, it was lower for surgically placed catheters (58.3%) compared with the 100% complication rate of percutaneously placed PD catheters. The importance of partial omentectomy is highlighted by the case in which a catheter initially was placed surgically without partial omentectomy but the patient continued to have complications until partial omentectomy performed, at which time the complications resolved. There were not enough data on the specific types of catheters used in this study to determine whether particular types of catheters affected the rate of catheter outflow complications, although other studies have remarked on complication rates of different catheters.13
In human patients receiving hemodialysis and PD, urea kinetics are used to assess adequacy of the dialysis dose. Generally, this involves calculating the fractional clearance of urea (Kt/V). This value requires measurements of urea clearance, both in dialysate and urine, as well as the daily urea generation rate. Because of the retrospective nature of this study, these values could not be obtained from existing patient data. URR has been described as an easier method to evaluate the effectiveness of dialysis in human patients receiving hemodialysis and PD.14 In at least 1 study, URR has been evaluated as a predictor of mortality in humans undergoing hemodialysis.15 In this study, a lower URR was associated with increased risk of mortality.15 In our study, we investigated the use of the URR as a predictor of survival in cats receiving PD.
In all measurements except for the 24-hour period, there were no significant differences in mean URR between survivors and nonsurvivors. Survivors had a lower URR than did nonsurvivors for the 24-hour period. The 12-hour mean URR for survivors was negative (ie, the mean BUN of survivors increased compared with predialysis results, whereas the mean BUN of nonsurvivors decreased).
It is difficult to interpret these results based on several aspects of data collection. Most cats in the study did not receive dialysis for 72 hours. As cats improved, PD was stopped, and many of the cats that did well did not receive PD for longer time periods. Thus, the longer time periods may have included animals that were more refractory to PD. Cats that fared poorly often were euthanized or died before 24 hours on PD, and no data could be collected for these cats. Cats that did well on PD and were weaned off dialysis before 24 hours also did not have URR calculated. Another problem that contributed to the low number of URR calculated was related to the method of measurement of BUN. All cats on PD had renal function monitored at least once daily with a clinical chemistry analyzer.f Additional reevaluation of renal function, electrolytes, and acid-base balance was performed on a NOVA cage-side analyzer,g which has an upper limit of measurement of 130 mg/dL for BUN. Although most cats had frequent reevaluation of renal function, many of them had a BUN concentration >130 mg/dL; therefore, URR could not be calculated for cats when BUN was measured by the NOVA cage-side analyzer. Because of these issues, not enough data could be collected to draw conclusions regarding the utility of the URR as a predictor of survival in this population. However, despite many issues with measurement of URR and urea kinetics in this population, additional study is recommended in patients receiving hemodialysis and PD.
In cases of acute kidney injury refractory to fluid therapy, both hemodialysis and PD are treatment options that can be considered. Hemodialysis requires specialized training, equipment, and water purification systems and is also not readily available to most veterinarians. PD is labor intensive, but does not require specialized equipment and may be performed in any clinic with adequate technical assistance and supervision. Patient personality also must be considered when deciding to perform PD because the patient must tolerate frequent manipulations. PD is an important therapeutic tool for mitigating clinical signs of uremia and giving the kidneys time to recover in cats with acute kidney injury when conventional therapy is no longer effective.
a SPSS for Windows version 13.1, SPSS Inc, Chicago, IL
b Kendall Inc, Mansfield, MA
c Cook, Spencer, IN
d Kendall Inc
e Ash Advantage, Medigroup Inc, Aurora, IL
f COBAS 6000, Roche Diagnostics, Indianapolis, IN
g Stat Profile, NOVA Biomedical Corporation, Waltham MA
The authors acknowledge Lori Lyn Price and Dr Louise Maranda for assistance in statistical analysis. This study was not supported by any funding or grants.