J Vet Intern Med 2012;26:275–281

Clinicopathological Variables Predicting Progression of Azotemia in Cats with Chronic Kidney Disease S. Chakrabarti, H.M. Syme, and J. Elliott Background: Chronic kidney disease (CKD) is common in geriatric cats, but often appears to be stable for long periods of time. Objectives: To describe CKD progression and identify risk factors for progression in newly diagnosed azotemic cats. Animals: A total of 213 cats with CKD (plasma creatinine concentration > 2 mg/dL, urine specific gravity < 1.035) were followed up until progression occurred or for at least 1 year; 132, 73, and 8 cats were in International Renal Interest Society (IRIS) stages 2, 3, and 4, respectively. Methods: Progression was defined as a 25% increase in plasma creatinine concentration. Logistic regression was used to assess variables at diagnosis that were associated with progression within 1 year. Changes in IRIS stage during followup also were described. Cases that remained in stages 2 or 3, but did not have renal function assessed in the last 60 days of life, were excluded from analysis of the proportion reaching stage 4. Results: Of the cats, 47% (101) progressed within 1 year of diagnosis. High plasma phosphate concentration and high urine protein-to-creatinine ratio (UPC) predicted progression in all cats. Low PCV and high UPC independently predicted progression in stage 2 cats, whereas higher plasma phosphate concentration predicted progression in stage 3 cats; 19% (18/94) of cats diagnosed in stage 2; and 63% (34/54) of cats diagnosed in stage 3 reached stage 4 before they died. Conclusions: Proteinuria, anemia, and hyperphosphatemia may reflect more progressive kidney disease. Alternatively, they may be markers for mechanisms of progression such as tubular protein overload, hypoxia, and nephrocalcinosis. Key words: Anemia; Hyperphosphatemia; Nephrology; Proteinuria.

hronic kidney disease (CKD) is common in geriatric cats, but often appears stable for long periods of time. Although several studies have evaluated survival in cats with CKD,1–6 few have documented changes in renal function. Survival studies also have been confounded by the difficulty of deducing cause of death because geriatric cats often have multiple pathologies and it is difficult to distinguish the cause of clinical deterioration. Furthermore, perceived quality of life has an influence on survival because the majority of cats are humanely euthanized rather than dying naturally. The first aim of this study was to describe progression of CKD in cats by evaluating changes in renal function over time. Plasma creatinine concentration is the most widely used marker of renal function, but can be affected by hydration status and muscle mass. The gold standard measure of renal function is glomerular filtration rate (GFR). However, GFR assessment consumes both time and resources. Because serial GFR measurements were not possible, changes in renal function were evaluated using the International Renal Interest Society (IRIS) staging system, which is

C

Abbreviations: ACE CKD GFR IRIS ROC UPC USG

angiotensin converting enzyme chronic kidney disease glomerular filtration rate International Renal Interest Society receiver operating characteristic urine protein to creatinine ratio urine specific gravity

based on plasma creatinine concentration (1.6–2.8 mg/ dL for stage 2, 2.9–5 mg/dL for stage 3, and > 5 mg/ dL for stage 4). Several markers have been associated with shorter survival in cats with CKD, including proteinuria,1 anemia,2 and hyperphosphatemia.4 The second aim of this study was to determine which clinicopathological variables could predict further deterioration of renal function in cats newly diagnosed with azotemic CKD.

Methods Case Selection

From the Department of Veterinary Basic Sciences, Royal Veterinary College, London, UK (Chakrabarti, Elliott); and the Department of Veterinary Clinical Sciences, Royal Veterinary College, Hatfield, UK (Syme). This work was carried out at the Royal Veterinary College, Royal College Street, Camden, London, UK. Corresponding author: S. Chakrabarti, Department of Veterinary Basic Sciences, Royal Veterinary College, Royal College Street, Camden, London, NW1 0TU; e-mail: [email protected].

Submitted August 19, 2011; Revised November 17, 2011; Accepted December 15, 2011. Copyright © 2012 by the American College of Veterinary Internal Medicine 10.1111/j.1939-1676.2011.00874.x

Cats with CKD were recruited from 1992 onward through geriatric cat clinics held at 2 first opinion London practices: the Peoples Dispensary for Sick Animals, Bow and Beaumont Sainsbury Animal’s Hospital, Camden. These cases were recruited for a variety of studies. Diagnosis was based on the concurrent findings of azotemia, which persisted for at least 2 weeks, and urine specific gravity (USG) that was < 1.035. Cases with shorter follow-up were only considered to have CKD if the history and clinical signs were compatible with chronic rather than acute kidney disease. Azotemia was defined as plasma creatinine concentration > 2 mg/dL (ie, above the laboratory reference range). Only cats with azotemic CKD recruited before June 2010 were included in the study and data were analyzed in June 2011.

276

Chakrabarti, Syme, and Elliott

Clinicopathological Data Blood and urine samples were obtained by jugular venipuncture and cystocentesis respectively at approximately 4-month intervals once cats were stabilized. These samples were analyzed for routine diagnostic purposes, obtaining plasma biochemical profiles and urine protein-to-creatinine ratio (UPC) from a commercial laboratory. Plasma biochemical profiles were obtained close to the time of sample collection, but UPC were obtained at a later date from urine samples that had been stored at 80°C. Urine protein was measured using a colorimetric pyrogallol red method and creatinine was measured using a colorimetric picric acid method. Sediment analysis was performed on all urine samples before storage and those with bacteriuria or > 5 white blood cells per high power field were cultured. Cats with urinary tract infections were treated with antibiotics for at least 3 weeks according to urine culture and sensitivity results. UPC from urine samples with gross hematuria or cats with urinary tract infections were not used in any analyses. Hypertension was diagnosed in cats with mean systolic blood pressure measurements > 170 mmHg on 2 consecutive visits or on 1 visit in association with hypertensive choroidoretinopathy. All hypertensive cats were treated with amlodipine at a daily dose of 0.625 mg, which was increased to 1.25 mg or subsequently 2.5 mg if blood pressure remained > 160 mmHg. All normotensive cats and hypertensive cats with normalized blood pressure were offered a commercial renal diet as standard care. The renal diets,a,b,c all clinical examinations, and biochemical tests were provided to clients free of charge. Other treatments were dispensed infrequently, according to the needs of the individual cat. These treatments included dietary phosphate binders, oral potassium supplementation and angiotensin converting enzyme (ACE) inhibitors, but no erythropoiesis-stimulating agents. Cats were routinely screened for hyperthyroidism using plasma total thyroxine concentrations and those that developed the disease at any time during follow-up were excluded from the study.

Risk Factors for Progression within 1 Year of Diagnosis Associations between clinicopathological variables at diagnosis and progression within 1 year were assessed. Progression was defined as a maximal increase in plasma creatinine concentration of at least 25% relative to the concentration at diagnosis. If there was evidence of dehydration at initial evaluation, baseline was considered to be the date at which normal hydration had been restored with fluid therapy, the azotemia had been stabilized, and fluid therapy had been discontinued for at least 1 week. The figure of 25% was chosen arbitrarily, but it was felt that smaller changes could be caused by lack of precision in the measurement of creatinine rather than actual progression of azotemia. In addition to a combined analysis using all azotemic cats, separate analyses were performed for cats diagnosed in stage 2 and those diagnosed in stage 3. Renal follow-up was defined as the interval between the date of diagnosis and the date of the last available plasma creatinine concentration. Nevertheless, cats often were seen at the clinics beyond the defined follow-up period. Cases that did not demonstrate progression but had renal follow-up < 1 year were excluded from the study.

Changes in IRIS Stage during Follow-Up Changes in IRIS stage during follow-up were described for the same population of cats, including those that were still alive or

eventually lost to follow-up. Only cats with renal function assessed in the last 60 days of life were considered to have been followed to death. Cats that were not followed to death are described separately. Of cats reaching a higher stage, the number that also demonstrated progression within 1 year of diagnosis was recorded.

Statistics Binary logistic regression was used to assess clinicopathological variables at diagnosis associated with progression within 1 year of diagnosis. Variables associated with progression in the univariate analyses were entered into backward selection multivariable models. USG was multiplied by 1,000 and UPC was multiplied by 10 to facilitate interpretation of the odds ratio. Receiver operating characteristic (ROC) curves were used to find optimal cut-points for variables predicting progression. Summary statistics are presented as median (25th, 75th percentile) and risk factors for progression are presented as odds ratio [95% confidence interval]. No attempt was made to impute missing data. All statistical analyses were performed using computer softwared with significance set at the 5% level.

Results From our database of over 2500 cases, 679 cats were diagnosed with CKD, but 212 cats were excluded due to concurrent hyperthyroidism. Two hundred and forty-five cats (118, 70, and 57 diagnosed in IRIS stages 2, 3, and 4, respectively) were excluded due to short biochemical follow-up. Although these cats did not demonstrate progression, they could not be considered stable because they did not have blood samples spanning a year from diagnosis. One hundred and sixty-four (67%) of the cats with short biochemical follow-up are known to have died or been euthanized within 1 year of diagnosis (including 54 of the stage 4 patients) and 81 (33%) were lost to follow-up. The cause of death was known in 45% (73/164) of the cats that died within 1 year of diagnosis and CKD was cited in 74% (54/73) of those cases. Eight further cases were excluded from the study because they showed progression, but this was not documented until after 1 year’s follow-up. Their status at 1 year was therefore uncertain. One additional cat with nephrotic syndrome was excluded because it had atypical disease pathology. These exclusions resulted in a study population of 213 cats including 119 males and 94 females. The majority (175) were nonpedigree domestic cats, but there were 14 Persian or Persian crosses, 14 Burmese, 4 Siamese or Siamese crosses, 4 British shorthairs, 1 Birman, and 1 Devon Rex. At baseline, there were 61 hypertensive cats, 118 normotensive cats, and 34 cats without blood pressure measurements. Thirty cats had urine cultured and 17 of these cultures yielded bacterial growth. At the time of data analysis, 135 cats had been euthanized, 33 had died and for 4 cats, although dead, it was not known whether the cat had died naturally or been euthanized. An additional 14 cats had been lost to follow-up and 27 were still alive.

Progression in Feline Chronic Kidney Disease

Risk Factors for Progression within 1 Year of Diagnosis One hundred and one cats were classed as having progressive disease, having demonstrated an increase in plasma creatinine concentration of at least 25% within 1 year of diagnosis. Plasma creatinine concentration increased by 51 (34, 83)% from 2.7 (2.4, 3.5) to 4.4 (3.5, 6.2) mg/dL in the progressive group. One hundred and twelve cats were considered stable having shown no such increase although they were followed for over a year. Plasma creatinine concentration increased by 7 (0, 16)% from 2.5 (2.3, 3.0) to 2.8 (2.5, 3.3) mg/dL in the stable group. Table 1 illustrates the variables at diagnosis considered in this study for cats in each group. Forty-two percent (56/132) of cats diagnosed in stage 2 progressed within 1 year of diagnosis, whereas the figures were 53% (39/73) and 75% (6/8) for cats diagnosed in stages 3 and 4, respectively. The clinicopathological variables associated with progression within 1 year of diagnosis in the combined univariate analysis of all azotemic cats included low body weight, PCV, USG and plasma albumin concentration, high UPC, plasma creatinine, urea and phosphate concentrations. These data are shown in Table 2. The multivariable model (Table 3) predicting progression of CKD within 1 year of diagnosis included UPC and plasma phosphate concentration. An increase in UPC of 0.1 was associated with a 24% increase in the risk of progression, and an increase in plasma inorganic phosphorus concentration of 1 mg/ dL was independently associated with a 41% increase in the risk of progression. Fifty-one (91%) of the progressive cats and 12 (16%) of the stable cats diagnosed in stage 2 reached stage 3 within 1 year of diagnosis. Table 4 shows the

277

clinicopathological variables associated with progression within 1 year of diagnosis in the univariate analysis of cats diagnosed at stage 2. Low PCV and high UPC remained in the multivariable model (Table 5). An increase in PCV of 1% was associated with a 10% reduction in the risk of progression and an increase in UPC of 0.1 was independently associated with a 23% increase in the risk of progression within 1 year of diagnosis. UPC was 0.23 (0.15, 0.40) in progressive cases and 0.13 (0.07, 0.24) in stable cases, whereas PCV was 33 (28, 36)% in progressive cases and 36 (33, 39)% in stable cases. Only 10 cats in the progressive group (18%) and 3 cats in the stable group (4%) were anemic (PCV < 27%). Twenty-seven (69%) of the progressive cats and 1 (3%) of the stable cats diagnosed in stage 3 reached stage 4 within 1 year of diagnosis. Table 6 shows the clinicopathological variables associated with progression within 1 year of diagnosis in the univariate analysis of stage 3 cats. The multivariable model included only high plasma phosphate concentration and an increase in plasma inorganic phosphorus concentration of 1 mg/dL was associated with a 43% increase in the risk of progression within 1 year of diagnosis. Exceeding the IRIS target for plasma inorganic phosphorus concentration in stage 3 cats (5 mg/dL)7 had a sensitivity of 74%, a specificity of 50%, a positive predictive value of 63%, and a negative predictive value of 63% for predicting progression within 1 year. Analysis of the ROC curve indicated no better cut-point.

Changes in IRIS Stage during Follow-Up At diagnosis, 62% (132) of the cats were in stage 2, whereas 34% (73) were in stage 3 and 4% (8) were in stage 4. Figure 1 shows the changes in IRIS stage for

Table 1. Summary statistics for clinicopathological variables at diagnosis in cats with chronic kidney disease. Data presented as median (25th, 75th percentile) or prevalence (%). Clinicopathological Variable Year of diagnosis Age (years) % Male Weight (kg) Plasma creatinine concentration (mg/dL) Plasma urea nitrogen concentration (mg/dL) Plasma inorganic phosphorus concentration (mg/dL) Plasma total calcium concentration (mg/dL) Plasma potassium concentration (mEq/L) Plasma cholesterol concentration (mg/dL) Plasma total protein concentration (g/dL) Plasma albumin concentration (g/dL) Plasma globulin concentration (g/dL) Packed cell volume (%) Urine protein to creatinine ratio Urine specific gravity % with urinary tract infection Systolic blood pressure (mmHg) % with systolic hypertension Renal follow-up (days) Median survival with 95% confidence interval (days)

Stable (n = 112) 2004 14 55 4.2 2.5 49 4.4 10.3 4.0 212 7.7 3.2 4.5 35 0.14 1.020 11 147 29 744 1021

(2000, 2007) (11, 16) (3.5, 4.9) (2.3, 3.0) (40, 60) (3.6, 5.2) (9.9, 10.7) (3.7, 4.4) (162, 247) (7.4, 8.2) (3.0, 3.4) (4.1, 4.9) (32, 38) (0.08, 0.24) (1.016, 1.024) (134, 164) (576, 1014) [861, 1181]

Progressive (n = 101) 2003 14 57 3.9 2.7 56 5.1 10.2 3.9 204 7.7 3.1 4.6 31 0.27 1.016 5 155 40 209 264

(1999, 2006) (11, 16) (3.2, 4.5) (2.4, 3.5) (45, 74) (4.0, 6.8) (9.6, 10.8) (3.6, 4.3) (166, 250) (7.3, 8.3) (2.9, 3.3) (4.1, 5.2) (25, 35) (0.17, 0.68) (1.014, 1.020) (132, 177) (99, 347) [186, 342]

278

Chakrabarti, Syme, and Elliott

Table 2. Univariate binary logistic regression analysis of clinicopathological variables at diagnosis associated with progression of chronic kidney disease in all azotemic cats. Only variables with P < .2 are presented here. Clinicopathological Variable Weight (kg) Plasma creatinine concentration (mg/dL) Plasma urea nitrogen concentration (mg/dL) Plasma inorganic phosphorus concentration (mg/dL) Plasma albumin concentration (g/dL) Packed cell volume (%) Urine protein to creatinine ratio 9 10 Urine specific gravity 9 1000 Presence of urinary tract infection Presence of systolic hypertension

Odds Ratio with 95% Confidence interval

N

P

0.68 [0.50, 0.93] 1.47 [1.06, 2.05]

194 213

.016 .022

1.02 [1.01, 1.03]

213

.008

1.34 [1.14, 1.57]

212