Grapes (Vitis spp.) are the largest fruit crop in the

Transcription

1 J Vet Intern Med 2005;19: Acute Renal Failure in Dogs After the Ingestion of Grapes or Raisins: A Retrospective Evaluation of 43 Dogs ( ) Paul A. Eubig, Melinda S. Brady, Sharon M. Gwaltney-Brant, Safdar A. Khan, Elisa M. Mazzaferro, and Carla M.K. Morrow A review of records from the AnTox database of the American Society for the Prevention of Cruelty to Animals Animal Poison Control Center identified 43 dogs that developed increased blood urea nitrogen concentration, serum creatinine concentration, or both as well as clinical signs after ingesting grapes, raisins, or both. Clinical findings, laboratory findings, histopathological findings, treatments performed, and outcome were evaluated. All dogs vomited, and lethargy, anorexia, and diarrhea were other common clinical signs. Decreased urine output, ataxia, or weakness were associated with a negative outcome. High calcium phosphorus product (Ca P), hyperphosphatemia, and hypercalcemia were present in 95%, 90%, and 62% of the dogs in which these variables were evaluated. Extremely high initial total calcium concentration, peak total calcium concentration, initial Ca P, and peak Ca P were negative prognostic indicators. Proximal renal tubular necrosis was the most consistent finding in dogs for which histopathology was evaluated. Fifty-three percent of the 43 dogs survived, with 15 of these 23 having a complete resolution of clinical signs and azotemia. Although the mechanism of renal injury from grapes and raisins remains unclear, the findings of this study contribute to an understanding of the clinical course of acute renal failure that can occur after ingestion of grapes or raisins in dogs. Key words: Calcium phosphorus product; Canine; Clinical toxicology; Hypercalcemia; Kidneys. Grapes (Vitis spp.) are the largest fruit crop in the world, 1 and production of grapes is second only to orange production in the United States. 2,3 Although native species exist worldwide, including Vitis labrusca, Vitis aestivalis, Vitis ripara, and Vitis rupestris in the United States, 4 Vitis vinifera is the main species of grape that is cultivated for the production of fresh fruit, raisins, wine, and juice. 1,4 A wide variety of cultivars and hybrids of V vinifera bear grapes for fresh fruit and wine production. The most common varieties grown in the United States are Thompson Seedless and Flame Seedless for fresh fruit and Chardonnay, Cabernet Sauvignon, Merlot, and Zinfandel for wine production. 5 During the off-season, fresh grapes are imported from Chile and Mexico. 6 Thompson Seedless grapes account for over 95% of US raisin production. 7,8 The grapes are sun-dried to produce natural seedless (dark) raisins or mechanically dried and then preserved with sulfur dioxide to produce golden raisins. 7,8 In 1999, veterinarians at the Animal Poison Control Center (APCC) of the American Society for the Prevention of Cruelty to Animals (ASPCA) received several reports in which seemingly healthy dogs developed acute renal failure (ARF) after ingestion of grapes or raisins. Later reports of grape and raisin toxicosis in dogs included a letter containing a brief summary of information about 10 symptomatic From the American Society for the Prevention of Cruelty to Animals (ASPCA) Animal Poison Control Center (Eubig, Gwaltney-Brant, Khan), Urbana, IL; the Departments of Veterinary Pathobiology (Brady) and Veterinary Biosciences (Morrow), College of Veterinary Medicine, University of Illinois, Urbana, IL; and the Denver Veterinary Specialists (Mazzaferro), Wheat Ridge Animal Hospital, Wheat Ridge, CO. Reprint requests: Paul A. Eubig, DVM, ASPCA Animal Poison Control Center, 1717 S Philo Rd, Suite 36, Urbana, IL 61802; peubig@apcc.aspca.org. Received September 9, 2004; Revised January 12, 2005; Accepted March 25, Copyright 2005 by the American College of Veterinary Internal Medicine /05/ /$3.00/0 dogs, 9 a brief report of raisin toxicosis in a Hungarian Vizsla, 10 a letter confirming that a similar syndrome had been identified in the United Kingdom, 11 and a case report describing the treatment of 4 dogs. 12 A similar problem in humans has not been identified. This study summarizes signalment, clinical signs, laboratory results, and response to treatment in 43 dogs that developed azotemia after ingestion of grapes or raisins. Materials and Methods Data Collection and Case Selection Historical and clinical information was obtained from the AnTox computerized medical record database of the APCC. The APCC is a 24-hour consultation service that receives phone calls from animal owners and practicing veterinarians throughout North America concerning incidents of animal poisoning. Information collected by the APCC on each animal includes signalment, source of exposure, amount of substance ingested, assurance of exposure (whether observed or suspected), types of clinical signs, time of onset and duration of clinical signs, laboratory findings, results of ancillary tests (eg, radiography, ultrasonography), histopathological findings, treatments performed, response to treatment, and final case outcome. The An- Tox database from 1987 through 2002 was searched for cases in which the APCC was contacted about ingestion of grapes or raisins by dogs. Dog owners, veterinarians, or both were contacted by phone or by facsimile to obtain updated information on the cases and copies of the medical records. The following criteria were required for inclusion in the study: (1) evidence of exposure (eg, witnessed exposure, grapes or raisins in the vomitus or stool, damaged packaging consistent with a dog ingesting the contents), (2) one or more clinical signs reported by the owner or the veterinarian, and (3) abnormally high blood urea nitrogen (BUN) or serum creatinine concentration after exposure. Cases were rejected if a dog also ingested other substances (eg, chocolate) or if a dog had evidence of preexisting health problems (eg, chronic renal failure) that might have contributed to the severity of the signs. The amount (ounces) of grapes or raisins ingested was recorded when available, converted to grams, and calculated as a dose (g/kg body weight). Information on clinical signs included the time from ingestion until the clinical sign was observed (time until onset) and duration of each sign (time until resolution). Clinical signs were not considered resolved if they still were present at the time of death. Upon reviewing the records, the authors found that anuria and oliguria

2 664 Eubig et al were subjective assessments rather than assessments based on measurement of urine output. As a result, anuria and oliguria were classified together as decreased urine output. Details recorded for 20 laboratory variables (BUN, creatinine, total calcium, phosphorus, calcium phosphorus product [Ca P], potassium, sodium, chloride, total CO 2 [TCO 2 ], anion gap, glucose, total protein, albumin, alkaline phosphatase [ALP], alanine aminotransferase [ALT], amylase, lipase, hematocrit, white blood cell count [WBC], platelet count) included initial values, peak and lowest abnormal values, and the times after exposure at which these variables were measured. In addition, the time after exposure at which the variable returned to the normal range was recorded for BUN, creatinine, total calcium, phosphorus, and Ca P. When urinalysis results were recorded in the record, results including urine specific gravity (USG), dipstick test results, and findings on evaluation of urine sediment were recorded. Only the first urinalysis performed on each dog was evaluated. Results of urine cultures and Leptospira titers were obtained if performed. Toxicological tests performed included quantitation of serum and urine from one dog for ethylene glycol. a Occasionally during the management of a patient, samples of grapes or raisins or tissue specimens were made available to the APCC for further testing. Samples were submitted for quantitation of liver and kidney 25-hydroxycholecalciferol [calcifediol; 25-(OH)D3] b in one dog and screening of liver for herbicides (23 agents) c in another dog. Additional testing included screening one sample of raisins for chlorinated hydrocarbon insecticides (15 agents), carbamate insecticides (15 agents), and organophosphorus insecticides (31 agents). c The same sample of raisins also was tested for the mycotoxins ochratoxin A, deoxynivalenol, fumonison B1, zearalenone, and aflatoxins. d In addition, 2 separate samples of grapes were tested for ochratoxin A. d Radiographic and ultrasonographic findings and results of histopathological evaluation were retrieved from the medical record when they had been performed. Specimens for histopathological evaluation were submitted to several different academic and commercial veterinary diagnostic laboratories by the attending clinicians. Treatments administered included fluid therapy, which was defined as the administration of IV fluids. Some dogs also received fluids by other routes, but the most accurate information obtained was about the IV route. Information recorded included the length of time until fluid therapy was initiated after exposure and the length of time that fluids were administered. The use of medications intended to increase the production of urine (eg, furosemide, dopamine, mannitol) was recorded. Other treatments noted included the use of other medications, whole-blood or plasma transfusions, and peritoneal dialysis or hemodialysis. Whether dogs died or were euthanized during the course of treatment or whether dogs survived was recorded. In addition, the numbers of days after exposure on which each dog was examined by the veterinarian and on which each dog had an evaluation of clinical chemistries were recorded. Statistical Analysis For statistical analysis, dogs that died and were euthanized were combined into 1 group and compared to dogs that survived. To reduce experiment-wise type I error, statistical analysis was performed only if data was available for 8 or more cases. A 2-sample independent t- test was used to compare differences in outcome with regard to age, weight, dose of grapes or raisins, time until initiation of fluid therapy, length of fluid therapy, initial abnormal values for the 20 laboratory variables, peak abnormal values for the 20 laboratory variables, lowest abnormal values for the 20 laboratory variables, and USG. Data were log-transformed if the assumptions of normal distribution were violated. A Mann-Whitney U-test was used when log transformation was insufficient to normalize the variables. To standardize for variability in measurement of chemistry and CBC results among different laboratories, each value was corrected based on the normal reference range reported for the value. For each value that was higher than the normal reference range, the corrected value was defined as the measured value divided by the value that defines the high end of the reference range for that laboratory. For each value that was decreased below the normal reference range, the corrected value was defined as the measured value divided by the value that defines the low end of the reference range for that laboratory. The corrected values were used for statistical analysis, but the measured values are reported because measured values are more recognizable to clinicians. Statistical comparisons excluded dogs with normal values, so differences in outcome were only compared for dogs with abnormal values. Also, dogs with unmeasurable values that exceeded the detection limit of the instrumentation were excluded from each comparison. A 2-tailed Fisher s test was used to compare differences in outcome with regard to sex, exposure to grapes versus raisins, use of medications intended to increase urine production (eg, furosemide, dopamine, mannitol), and presence versus absence of each clinical sign. Odds ratios were determined to quantify the degree of association among variables and are reported with 95% confidence intervals (CI). Pearson s r was determined to evaluate the correlation among time until death and age, weight, or dose. For all analyses, a value of P.05 was considered significant. Analyses were conducted by commercially available statistical software. e,f Results Forty-three of 132 cases that were reviewed met the inclusion criteria. The reasons for which cases were rejected included poor historical information or poor prior health (41 dogs); exposure to grapes/raisins for which neither signs nor azotemia developed (33), exposure to grapes/raisins with the development of signs but no azotemia (14), and possible involvement of co-ingestants that could confound the evaluation of the clinical course (15). Complete copies of the medical records of 30 of 43 (70%) dogs included in the study were obtained and reviewed. Complete copies of the results of laboratory testing were obtained for another 9 (of 43; 21%) dogs. Clinical information for these dogs was recorded from incomplete copies of the medical records and from the AnTox database information collected at the time the APCC was originally contacted about the case. For the remaining 4 dogs in the study, a combination of information from the AnTox database and from case summaries provided by the veterinarians who treated the dogs was reviewed. Signalment The most commonly involved breed was Labrador Retriever (13 dogs) followed by Golden Retriever (4) and mixed breed (4), with 17 other breeds comprising the remaining 22 dogs. The median age was 4.0 years (range, years). No significant difference in age between dogs that survived and those that died or were euthanized was observed (P.077), nor was a correlation found between the times until death and the ages of the dogs that did not survive (P.62). Median weight was 25 kg (range, 2 48 kg). Weights were not available for 2 dogs. No significant difference in the weights was identified between dogs that survived and those that died or were euthanized (P.31), nor was there a correlation between the times until death and the weights of the dogs that did not survive (P.88). Twenty-seven (62.8%) of the dogs were male (1 intact, 25 neutered), and 16 (37.2%) were female (2 intact, 12 spayed). One male and 2 females were of unknown reproductive status. Because of the relatively low numbers

3 Grapes in Dogs 665 Table 1. Clinical signs reported in 43 dogs after the ingestion of grapes or raisins. Sign % Dogs Exhibiting Sign (No. of dogs) Median Time Until Onset of Sign (range, in days) a Median Time Until Resolution of Sign (range, in days) b,c % (No. of dogs exhibiting sign) That Survived Odds Ratio (95% confidence interval) d P Value Vomiting 100 (43/43) 0.5 (0 2) e 5 (0.6 14) 53 (23/43) ND ND Lethargy 77 (33/43) 1.1 (0.3 10) 6 (0.6 9) 55 (18/33) (0.29, 4.95) Anorexia 72 (31/43) 0.8 (0 3) e 7 (1.1 19) 61 (19/31) (0.78, 12.80) Diarrhea 51 (22/43) 0.9 (0.2 7) 2 (0.3 13) 41 (9/22) (0.10, 1.20) Decreased 49 (21/43) 3 (0.8 7) 1.5 (0.2 6) 24 (5/21) urine output f (0.016, 0.30) Abdominal 28 (12/43) 3 (0.8 8) 1.3 (0.9 3) 50 (6/12) pain (0.22, 3.13) Ataxia f 23 (10/43) 2 (0.6 6) 6 10 (1/10) (0.006, 0.50) Weakness f 19 (8/43) 2 (0.8 6) (1/8) (0.0093, 0.76).016 ND, not determined. a Time until onset of sign, defined as the time from ingestion until the sign began. b Time until resolution of sign, defined as the duration of each sign once present. c Includes dogs that survived and dogs in which the sign was noted to have resolved prior to death. d Indicates odds ratio for survival for dogs that exhibited the sign. e Value of 0 indicates that the sign was reported as beginning immediately after ingestion. f Significant at P.05. of intact dogs, the evaluation of sex differences between dogs that survived and those that did not survive was limited. When only the neutered dogs were evaluated, spayed female dogs may have been more likely to survive than were neutered male dogs (P.017; odds ratio, 7.50 [CI: 1.35, 41.70]). However, when all dogs were evaluated, no significant difference in survival between the sexes was observed (P.056; odds ratio, 4.36 [CI: 0.95, 22.85]). Exposure Twenty-eight dogs ingested raisins, 13 ingested grapes, and 2 dogs ingested both. The types of raisins ingested were mainly natural seedless (dark) raisins, but golden seedless raisins were ingested by at least 1 dog. In 1 dog, the raisins were described as organic. Estimates of the amount of raisins ingested were available for 24 of the dogs that ingested only raisins. The median amount ingested was 448 g (16 oz), with a range of 42 to 896 g ( oz) reported. The estimated dosage of raisins ranged from 2.8 to 36.4 g/kg ( oz/kg), with a median dosage of 19.6 g/kg (0.7 oz/ kg). No statistically significant difference was identified in the dosages of raisins ingested by dogs that survived and those that died or were euthanized (P.46), and no correlation was found between the times until death and the dosages of raisins ingested in dogs that did not survive (P.86). The grapes ingested included those from vines on the owners properties, grapes purchased from grocery stores, and wine grapes purchased from other sources. Identification of the cultivars of grapes was not available. In 2 cases, the grapes were noted to have visible evidence of decay. Estimates of the amount of grapes ingested were available for 5 dogs. The amounts ingested varied, with 1 dog ingesting 10 to 12 grapes and the other 4 dogs ingesting from 448 to 1,344 g (16 48 oz). The dosages for the latter 4 dogs were 19.6, 30.8, 50.4, and g/kg (0.7, 1.1, 1.8, and 5.3 oz/kg), respectively. As a result of small sample size, analyses were not performed with regard to the dosages of grapes. No significant difference in outcome was identified between dogs that ingested raisins and dogs that ingested grapes (P.52; odds ratio, 0.62 [CI: 0.16, 2.39]). Clinical Signs Information was collected on reported clinical signs. Table 1 includes the most common signs (reported in 8 dogs). Vomiting was the most commonly reported clinical sign, occurring in 100% of affected dogs. Signs significantly associated with a negative outcome included ataxia (P.027), decreased urine output (P.0002), and weakness (P.016). Although anuria and oliguria were classified together as decreased urine output due to the subjective nature of these assessments, 15 of 17 dogs reported to be anuric died or were euthanized. Statistical analysis could not be performed on dogs that vomited because no comparison group of dogs that did not vomit was available. Signs reported in 7 dogs include hypertension, dehydration, edema, hypothermia, trembling, polydipsia, polyuria, ptyalism, and seizures. The earliest and most consistent clinical sign in these dogs was vomiting, which occurred within the first 24 hours after ingestion in the majority of dogs (35/43, 81%). Grapes or swollen raisins often were observed in the vomitus. Anorexia (20/43, 47%), lethargy (15/43, 35%), and diarrhea (12/43, 28%) often developed within the first 24 hours, after the onset of vomiting. In some instances, grapes or swollen raisins also were present in diarrheal fluid. Within 24 to 48 hours after ingestion, vomiting developed in the remaining

4 666 Eubig et al Table 2. Clinicopathologic findings in 43 dogs after the ingestion of grapes or raisins. a Variable (No. of dogs evaluated) Timing of Measurement Day Postexposure, Median (range) Value, Median (range) Proportion (%) of Dogs for Which Value Reported b (absolute) Proportion (%) of Dogs With Abnormal Results c (absolute) 2 (0.9 8) 2.5 (0.9 10) BUN (n 43) 2 (0.8 8) 99 mg/dl (3 196) 100 (43/43) Peak 4 (0.8 8) 128 mg/dl (28 258) 98 (42/43) Creatinine (n 43) 2 (0.8 8) 9.9 mg/dl ( ) 100 (43/43) Peak 3 (0.8 8) 10.4 mg/dl ( ) 100 (43/43) Total calcium (n 40) d 12.3 mg/dl ( ) 93 (40/43) Peak d 2 (0.9 9) 13.5 mg/dl ( ) 62.5 (25/40) Phosphorus (n 41) 3 (0.9 8) 10.6 mg/dl ( ) 95 (41/43) Peak 3.5 (1.1 8) 13.0 mg/dl ( ) 90 (37/41) Ca P(n 40) d 139 (42 426) 93 (40/43) Peak d,e 3 (0.9 10) 151 (66 426) 95 (38/40) Potassium (n 36) d 2 (1 8) 5.05 meq/l ( ) 84 (36/43) Peak d Low d 3 (1.9 8) 7.5 (3 31) 6.3 meq/l ( ) 3.35 meq/l ( ) 42 (15/36) 44 (16/36) Sodium (n 30) 2 (1 8) 147 meq/l ( ) 70 (30/43) Low 6.5 (2 18) 136 meq/l ( ) 33 (10/30) Chloride (n 28) 2.5 (1.1 8) meq/l (80 117) 65 (28/43) Low 3 (1.1 12) 98 meq/l (80 109) 68 (19/28) Total CO 2 (n 17) Anion gap (n 16) d 3 (1.4 8) 3.5 (1.4 8) 19 meq/l (9 26) 32.5 (8 44) 40 (17/43) 37 (16/43) Peak 3 (2 8) 34 (32 44) 69 (11/16) Glucose (n 38) 2 (0.8 8) 109 mg/dl (75 238) 88 (38/43) Peak 3.5 (0.8 54) mg/dl ( ) 47 (18/38) Total Protein (n 37) 2 (0.9 8) 6.4 g/dl ( ) 86 (37/43) Low 4 (2 11) 3.8 g/dl ( ) 22 (8/37) Albumin (n 38) 2 (0.9 20) 3.5 g/dl ( ) 88 (38/43) ALP (n 38) 2 (0.9 8) 81 U/L (22 348) 88 (38/43) Peak 5 (1 13) 285 U/L (95 1,242) 24 (9/38) ALT (n 40) 2 (0.8 8) 102 U/L (32 385) 93 (40/43) Peak 3 (0.8 15) U/L (90 491) 65 (26/40) Amylase (n 38) Lipase (n 29) 2 (0.7 8) 3 (0.7 8) 649 U/L (212 3,524) 458 U/L (49 3,039) 88 (38/43) 67 (29/43) Peak 3 (1.1 8) 1,008 U/L (334 3,039) 41 (12/29) Hematocrit (n 37) 2 (0.8 8) 42.2% ( ) 86 (37/43) Low 4 (1.9 18) 31% ( ) 41 (15/37) WBC (n 33) Platelet Count (n 30) 2 (0.8 8) 2 (0.8 13) / L (3.8 16) / L (11 522) 77 (33/43) 70 (30/43) Low 4.5 (1.8 19) / L (6 198) 27 (8/30) BUN, blood urea nitrogen; Ca P, calcium phosphorus; ALP, alkaline phosphatase; ALT, alanine aminotransferase; WBC, white blood cell count. a Not all parameters are available for each of the 43 dogs. b For initial values, this was calculated by dividing the number of dogs for which the parameter was measured by 43 dogs. c For peak or low values, this was calculated by dividing the number of dogs that deviated from the normal range by the number of dogs for which the parameter was measured. d Significant difference between dogs that survived and dogs that died or were euthanized at P.05. See Table 3 for further details. e Ca P considered abnormal if dogs, and lethargy developed in an additional 11 dogs (26%). Although in some dogs signs such as decreased urine output, abdominal pain, ataxia, and weakness began as early as the day of exposure, it generally took up to 5 days for the majority of dogs to exhibit these clinical signs. Clinicopathologic Data Information on 20 laboratory variables in the 43 dogs is reported in Table 2. Values were included if findings were available for 8 dogs. Ninety-five percent (38/40) of the dogs evaluated in this series had high Ca P, 90% (37/41) were hyperphosphatemic, and 62.5% (25/40) were hypercalcemic. Significant differences in initial (P.044) and peak (P.047) serum total calcium concentration, initial (P.011) and peak (P.005) Ca P, initial (P.003) and peak (P.042) and low (P.03) serum potassium concentration, and initial TCO 2 (P.029) were identified when values for dogs that survived were compared to those of dogs that died or were euthanized. Table 3 compares the values of the laboratory variables that were significantly different between dogs that survived and those that died or were euthanized. Although hyperkalemia and hypokalemia

5 Grapes in Dogs 667 Table 3. Comparisons of values of laboratory variables in which statistically significant differences were detected between dogs that survived and dogs that died or were euthanized. Variable (No. of dogs evaluated) a Median Value (range) of Dogs That Survived Proportion (%) of Dogs That Survived (absolute) Median Value (range) of Dogs That Died or Were Euthanized Proportion (%) of Dogs That Died or Were Euthanized (absolute) Total calcium, (n 40) 11.7 mg/dl ( ) 57.5 (23/40) 13.5 mg/dl ( ) 42.5 (17/40).044 Peak (n 25) 12.7 mg/dl ( ) 48 (12/25) 14.1 mg/dl ( ) 52 (13/25).047 Ca P, (n 38) 117 (42 179) 60.5 (23/38) 159 (66 426) 39.5 (15/38).011 Peak (n 36) 134 (73 179) 58 (21/36) 181 (66 426) 42 (15/36).005 Potassium, (n 36) 4.6 meq/l ( ) 58 (21/36) 5.4 meq/l ( ) 42 (15/36).003 Peak (n 15) 5.8 meq/l ( ) 27 (4/15) 6.4 meq/l ( ) 73 (11/15).042 Low (n 16) 3.2 meq/l ( ) 87.5 (14/16) 3.8 meq/l ( ) 12.5 (2/16).03 Total CO 2, (n 17) 21 meq/l (9 26) 65 (11/17) 15 meq/l (9 20) 35 (6/17).029 P Value a For initial values, n number of dogs for which this parameter was evaluated. For peak or low values, n number of dogs for which this parameter was abnormal. developed in similar numbers of dogs (15/36 or 42% and 16/36 or 44%), 11 of 15 of the hyperkalemic dogs died or were euthanized, whereas only 2 of 16 of the hypokalemic dogs died. Serum creatinine and BUN concentrations were available within 24 hours of ingestion for 3 dogs. Serum creatinine concentrations for all 3 dogs were 2.5 to 3.3 times above normal, whereas BUN concentrations were less consistently abnormal in these 3 dogs and ranged from normal to 1.7 times above normal. Serum phosphorus and total calcium concentrations were available for 2 of 3 dogs, and the phosphorus concentrations were 1.2 and 1.5 times above normal, whereas the total calcium concentrations were only slightly above normal. The Ca P was high for both dogs, with values of 92 and 130, respectively. Eight dogs were evaluated for the first time within 24 to 48 hours after ingestion. These dogs had high serum creatinine (8/8), BUN (7/8), phosphorus (4/8), and calcium (2/8) concentrations and high Ca P (6/8). It was not until 48 to 72 hours after exposure that serum calcium concentrations were clearly increased, with 12 of 13 dogs first evaluated during this time period exhibiting hypercalcemia. Urinalysis Urinalysis results were available for 29 dogs, with sediment examination having been performed in 24 dogs. Urinalysis was performed a median of 3 (range, 0.9 to 9) days after exposure. USG was in 97% (28/29) of dogs, with 41% (12/29) isosthenuric and 21% (6/29) hyposthenuric. No significant difference was found in USG between dogs that survived and those that died or were euthanized (P.89). Urinalysis was performed after fluid therapy was initiated in 13 of 29 dogs and before treatment in 2 dogs. The timing of urinalysis in relation to the initiation of fluid therapy could not be reliably determined in the other 14 dogs. Median ph was 6.9 (range, ) in the 26 dogs for which it was reported. Proteinuria was identified in 56% (15/27) dogs, in which trace proteinuria was present in 8 of 27 dogs and more than trace in 7 of 27 dogs. Glucosuria was present in 50% (12/24) of dogs, with trace amounts in 7 of 24 dogs and larger amounts in 5 of 24 dogs. Concurrent blood glucose concentrations were available for 3 of these 5 dogs, with 2 having normal blood glucose concentrations and 1 dog having mild hyperglycemia (138 mg/dl; reference range mg/dl). No dogs had ketonuria or bilirubinuria. Microscopic hematuria was identified in 58% (14/24) of dogs. Six of these had 5 red blood cells/high power field, with concurrent proteinuria in 5 of 6 dogs. WBCs were observed in the urine of 67% (16/24) of dogs, but only 1 dog had 10 WBC/hpf. Crystalluria was reported in 33% (8/24) of dogs, with the following types: magnesium ammonium phosphate crystals (4 dogs), calcium oxalate monohydrate crystals (1), calcium oxalate monohydrate and dihydrate crystals (1), urate crystals (1), and amorphous crystals (1). Cylindruria was reported in 21% (5/24) of dogs, and 3 of these dogs had granular casts and 2 had hyaline casts. Cylindruria was detected in all 5 dogs within 72 hours of grape or raisin ingestion. Microbiology and Serology Results of urine cultures were recorded for 4 dogs. No growth was identified in 3 dogs. Organisms associated with in-dwelling urinary catheterization were cultured from the remaining dog. Microscopic agglutination titers for Leptospira serovars were performed on serum samples from 8 dogs, and results either were negative or were suggestive only of prior vaccination or exposure. Toxicology One dog had serum and urine evaluated for ethylene glycol 3 days after signs began, but no ethylene glycol was detected. Results of quantitation of 25-(OH)D3 performed on renal and hepatic specimens collected 7 days after ingestion in 1 dog were unremarkable. Evaluation of a specimen of liver from 1 dog collected 9 days after ingestion and submitted for a herbicide screen indicated no detectable herbicides. No evidence of chlorinated hydrocarbon, car-

6 668 Eubig et al bamate, or organophosphorus insecticides was found upon screening of raisins that remained after ingestion. No mycotoxins were detected in the same sample of raisins, and no evidence of ochratoxin A was found in 2 separate samples of grapes that were evaluated. g Radiographic and Ultrasonographic Findings Results of abdominal radiography were normal in 19 of 24 (79%) dogs evaluated, and thoracic radiography was normal in 1 of 5 dogs evaluated. Renal size and shape were normal in most dogs evaluated. Two dogs had abdominal effusion and pleural effusion, and 3 had pulmonary edema. Fluid therapy had already been started in all 3 dogs with effusions or pulmonary edema at the time radiographs were taken. Pulmonary soft tissue mineralization was evident 7 days after exposure in another dog, with a peak Ca Pof 179. Abdominal ultrasonography identified abnormalities in 11 of 13 (85%) dogs, with 7 of 13 (54%) having changes in the kidneys, including renomegaly (2 dogs), hyperechoic renal cortices (1), renomegaly and hyperechoic renal cortices (1), renomegaly and pelvic dilatation (1), pelvic dilatation (1), and unspecified change in renal echogenicity (1). Two dogs had mixed echogenic changes in the pancreatic parenchyma, and 1 dog had hyperechoic peripancreatic mesentery. Histopathology Renal histopathology was evaluated in 16 dogs, and other tissues, including liver, heart, and lung, were evaluated in 10 of these dogs. The most consistent lesion present in biopsy or postmortem samples from 15 of 16 dogs was moderate to severe, diffuse renal tubular degeneration, especially in the proximal tubules. Proteinaceous and cellular debris often was present in dilated tubule lumens. Other lesions included mineralization of necrotic epithelial cells or tubular basement membranes (9 dogs), tubular epithelial cell regeneration (6), mineralization of glomerular basement membranes (1), diffuse glomerular ischemia (1), glomerular congestion (1), and interstitial nephritis (1). No tubular degeneration was noted in 1 dog, but chronic interstitial nephritis and proliferative glomerulonephritis were observed 14 days after exposure to grapes. This same dog had bacterial growth on urine culture and had a positive serum titer for Leptospira grippotyphosa before renal biopsies were obtained. Mineralization of several extrarenal tissues, including gastric mucosa, myocardium, lung, and blood vessel walls, was reported in 4 of 16 dogs. One dog had acute to subacute, multifocal pancreatic and fat necrosis. Other lesions reported included congestion or edema of liver (8 dogs), lungs (3), spleen (2), pancreas (1), and abdominal lymph nodes (1), as well as occasional fibrinoid vasculitis or vascular necrosis (3). Treatment All but 1 dog received IV fluid therapy. The median number of days after exposure until fluid therapy was initiated was 2.8 days (range, days). Fluid therapy was typically initiated on the day the dogs were presented, but in 6 dogs, a delay of 1 to 3 days occurred before fluids were given. The length of time until fluid therapy was initiated was not significantly different between dogs that survived and those that did not (P.73). The length of time fluids were administered was significantly longer (P.001) in dogs that survived (median, 6.0 days [range, 0 14 days]) than in dogs that died or were euthanized (median, 2.5 days [range, days]). Forty-two percent (18/43) of dogs received furosemide, dopamine, mannitol, or some combination of these drugs. No significant difference in outcome was identified when dogs that received these medications were compared to those that did not (P.76; odds ratio, 1.27 [CI: 0.38, 4.29]). Other types of medications administered included gastrointestinal protectants, antiemetics and motility modifiers, phosphate binders, antibacterial agents, corticosteroids, amlodipine, and heparin. Four dogs received wholeblood transfusions, and 2 of these dogs also received fresh frozen plasma. Two dogs received peritoneal dialysis, and of these 2 dogs, 1 recovered and 1 died. One dog received hemodialysis for a total of 8 treatments over 16 days and recovered. Outcome Outcomes for the 43 dogs included survival in 23 dogs (53%), death in 5 dogs (12%), and euthanasia in 15 dogs (35%). The median time from exposure until death or euthanasia was 6 days (range, 1 11 days). Clinical signs resolved for all 23 dogs that survived. Of the 23 dogs, 15 had a documented return to within-normal ranges of BUN and serum creatinine concentrations. Eight dogs were azotemic at their last recorded evaluation 7 to 104 days (median, 16 days) after ingestion of grapes or raisins. In dogs that survived, BUN returned to normal in 18 of 22 dogs in a median time of 15.5 days (range, 3 59 days). The last recorded BUN ranged from 34 to 58 mg/dl in the 4 dogs with persistently increased BUN concentration. Serum creatinine concentration returned to normal in 18 of 23 dogs in a median time of 30 days (range, days). The last recorded serum creatinine concentration ranged from 1.6 to 2.4 mg/dl in the 5 dogs with persistently high serum creatinine concentrations. Serum total calcium concentrations returned to normal in 11 of 12 dogs in a median time of 11 days (range, 2 51 days). Serum phosphorus concentrations returned to normal in 18 of 20 dogs in a median time of 8 days (range, 2 19 days). Ca P returned to normal in 17 of 21 dogs in a median time of 11 days (range, 2 19 days). Discussion The 43 dogs in this report had clinical signs and some degree of azotemia after the ingestion of grapes, raisins, or some combination of both. All dogs had high serum creatinine concentrations, and most dogs had marked increases in both BUN and serum creatinine concentrations. Every dog in this study may not have been in ARF, because variable timing of laboratory testing and infrequent determination of USG before fluid therapy make it hard to distinguish between prerenal and renal azotemia in some of these dogs. The dog with chronic interstitial nephritis and prolif-

7 Grapes in Dogs 669 erative glomerulonephritis likely had preexisting renal insufficiency. Five months before the ingestion of grapes the dog had normal values for BUN and creatinine. Dogs in this study usually began vomiting within 24 hours of ingestion; vomiting was followed by development of anorexia, lethargy, and diarrhea. Serum creatinine and phosphorus concentrations and Ca P increased soon after exposure, whereas BUN tended to increase 24 hours later, and serum calcium concentrations increased 48 to 72 hours after exposure. Vomiting was a consistent finding and was observed more frequently in dogs of this report (100%) than in those of another study of dogs with ARF due to other causes (64%). h The finding that dogs with decreased urine output were 14 times less likely to survive is not unexpected and resembles findings in another study of ARF in dogs. 13 Dogs with ataxia and weakness were 18 and 12 times less likely to survive, respectively. Neuromuscular signs can be seen in uremic patients 14,15 and more commonly develop during acute uremia. 15,16 The majority of dogs in this study had high Ca P, with median values that were higher for dogs that died or were euthanized than for those that survived. Alterations in Ca P may have contributed to the progression of renal injury. Hypercalcemia promotes soft tissue mineralization, which is most pronounced when Ca P exceeds 50 to ,18 Tissues damaged by mineralization include the renal tubular epithelium and basement membrane, atrial myofibers and subendocardial tissue, pulmonary alveolar walls, gastric mucosa, and vascular tunica media Dogs in this study with hypercalcemia, high Ca P, or both may have poorer outcomes than those that did not share such values because they are more likely to experience progressive renal dysfunction and to develop soft tissue mineralization. Alternatively, higher serum total calcium concentration and Ca P may simply be an indication of more severe renal damage. Measurement of serum ionized calcium and parathyroid hormone concentrations and urinary calcium concentrations in symptomatic dogs may help elucidate the roles of calcium and phosphorus in dogs that develop ARF after the ingestion of grapes or raisins. Hyperkalemia in renal failure most commonly is associated with oliguria. 20 Not all dogs in this study were reported to have decreased urine output at the time hyperkalemia was identified, but 13 of 15 hyperkalemic dogs also were reported to have decreased urine output at some point in their clinical course. Hypokalemia is more commonly seen with polyuric renal failure or as a result of fluid therapy. 20 A low TCO 2 concentration, which indicates the presence of metabolic acidosis, was another negative prognostic indicator in dogs developing ARF after ingestion of grapes or raisins. Of the 40% of dogs that had TCO 2 evaluated, the median value in dogs that died or were euthanized was significantly lower than that of those that survived. Metabolic acidosis can cause deleterious cardiovascular, respiratory, neurologic, and metabolic effects 21,22 that may have contributed to mortality in these dogs. Metabolic acidosis often is associated with oliguric renal failure in dogs. 14 Not all of the dogs with low TCO 2 concentrations that died or were euthanized were reported to have decreased urine output, but all developed hypercalcemia and high Ca P, which may have contributed to the progression of ARF. Fluid therapy was initiated before urine was submitted for urinalysis in many of the dogs, and this probably confounded evaluation of USG in relation to outcome. One dog had pancreatitis confirmed by histopathology, and 2 other dogs had ultrasonographic changes suggestive of pancreatitis. 23 Pancreatic and peripancreatic edema can develop due to overhydration 24 and hypoalbuminemia, 25 which may account for some of the changes seen on ultrasound. Ultrasonographic findings such as renal pelvic dilatation, splenic congestion, and abdominal effusion seen in other dogs in this study indicate that overhydration was a concern during their treatment. Amylase and lipase concentrations were not increased in the 2 dogs evaluated by ultrasound and were not assessed before death in the dog that had histopathological evidence of pancreatitis, but high amylase and lipase concentrations are not consistent findings in dogs with acute pancreatitis. 26 The association of suspected pancreatitis with ARF in these dogs is uncertain. Acute pancreatitis has been reported in human patients on maintenance dialysis and after renal transplantation 27 but is rarely reported in acutely uremic patients. Hypercalcemia has been associated with development of pancreatitis in dogs 28 and in humans. 29 The 3 dogs had peak increases in serum total calcium concentration up to 1.2 times the normal range, and the highest recorded Ca P was 136. The report by Neuman 28 indicates that hypercalcemia does not necessarily need to be prolonged to induce pancreatitis. Dogs with ARF after grape or raisin ingestion should be monitored for evidence of pancreatitis, especially if clinical status worsens. There often was a delay in the initial evaluation of the dogs in this study, presumably because most pet owners and their veterinarians suspected that the dogs were only experiencing mild gastritis. A delay in evaluating a symptomatic dog after the ingestion of grapes or raisins may be detrimental, because serum creatinine concentrations may increase within the first 24 hours after exposure. Although the length of time until fluid therapy was initiated was not found to correlate with outcome in these dogs, the effect of initiating fluid diuresis before the onset of clinical signs could not be evaluated. In these dogs, the duration of fluid administration was important. Dogs that survived were more likely to have received fluids for a longer period of time. Dogs that did poorly may have died or been euthanized due to rapid progression of their illness relatively soon after exposure, whereas dogs that survived may have had milder or more slowly progressive disease and continued to receive fluid therapy while they recovered. Alternatively, dogs may have survived because they were allowed to receive fluids for a longer period of time. The latter interpretation implies that some dogs may have been euthanized prematurely. Review of the individual cases, however, indicated that euthanasia usually was performed due to progressive disease or failure to respond clinically. Three other studies of ARF in dogs found that the time of hospitalization (during which IV fluids presumably would be given) was significantly longer for dogs that survived until discharge. 13,30,h Unfortunately, these findings do not allow additional clarification of the importance of fluid ther-

8 670 Eubig et al apy in the treatment of dogs that develop ARF after ingestion of grapes or raisins. The outcome of the dogs in this series is somewhat better than observed in other studies of ARF in dogs, 13,30,h but the short-term prognosis still must be very guarded. Fifty-three percent of the dogs that developed ARF after the ingestion of grapes or raisins recovered. This result is compared to 38% of dogs surviving in a study of hospital-acquired ARF, 13 20% of dogs surviving in another study, h and 44% of dogs surviving in yet another study. 30 The latter 2 studies contained dogs with various causes of ARF, including ethylene glycol, which has a grave prognosis once renal failure develops. 31 Dogs that developed ARF and survived after ingestion of grapes or raisins had a fair long-term prognosis, with 61% (15 dogs) having documented resolution of clinical signs and azotemia. BUN and serum creatinine concentrations returned to normal more slowly than did serum calcium and phosphorus concentrations and Ca P. The other 39% (8 dogs) that recovered had resolution of clinical signs but residual azotemia indicating prolonged recovery or permanent reduction in renal function. These 8 dogs had no prior history of clinical signs such as polyuria or polydipsia, but prior serum chemistry and urine evaluations were not performed on these dogs, and it can not be proven that preexisting renal disease was not present. The toxic principle that causes ARF after ingestion of grapes or raisins in dogs is unknown. Acute tubular necrosis, especially in the proximal tubules, was the most consistent histopathological finding in the 16 dogs in which renal tissue was evaluated. Acute renal failure generally is caused by toxic or ischemic injury. 32,33 Some chemical component of grapes and raisins may either have caused direct renal tubular damage or caused renal ischemia in the dogs in this study. The renal cortex has a high rate of oxygen consumption (making it susceptible to ischemic injury) and a relatively large endothelial surface area compared to other organs. Substantial delivery of a toxic agent to the kidneys can occur as a result of these factors. In addition, active and passive tubular transport may concentrate toxic substances in the tubular epithelium or renal interstitium. Multiple tubular epithelial enzyme systems exist, especially in the proximal tubules. These enzyme systems are better known for detoxifying substances, but activation of compounds to more toxic forms by the tubular epithelium, as is seen in acetaminophen toxicosis in humans, 34 can occur. Biotransformation of a compound to a more toxic intermediate by other metabolic enzyme systems elsewhere in the body also may occur, with subsequent transport of the activated compound to the kidneys. Ischemic tubular injury potentially can be distinguished from nephrotoxic tubular injury based on histopathological changes. The basement membranes may be disrupted by ischemic damage but tend to remain intact after a toxic insult. 35 Lack of basement membrane disruption, however, does not exclude the possibility of ischemia. As the clinical course progresses, compensatory changes and regeneration make it more difficult to distinguish among the causes of acute tubular necrosis. The relative severity of tubular damage also can confound interpretation of the changes. It is difficult to determine from the available histopathological results whether or not ischemic tubular necrosis developed in the dogs in this study because samples usually were obtained after several days of therapy. Tubular epithelial regeneration was identified in 38% of dogs in which histopathology was evaluated. Generally, tubular recovery can occur if the tubular basement membranes remain intact and tubular epithelial cells remain viable, 32 and this finding indicates that at least partial recovery of renal function can occur in dogs when ARF develops after ingestion of grapes or raisins. Acute renal failure does not consistently develop in dogs that ingest grapes or raisins in doses similar to or higher than those that dogs in this study ingested. Over half of the cases that were rejected for inclusion in this study were rejected because either no abnormal clinical signs developed or because abnormal clinical signs were present without azotemia. In dogs in this study, outcome was not influenced by the dosage of raisins ingested. Dosages of as little as 2.8 g/kg (0.1 oz/kg) of raisins and 19.6 g/kg (0.7 oz/kg) of grapes resulted in ARF. Generally, toxic agents exhibit a dose-response relationship in which a changing degree of response is seen as a changing amount of a toxic agent is administered. 36 Given the lack of an apparent dose-response relationship, the toxic principle may be an intrinsic compound that is present in varying amounts in Vitis spp. depending on environmental circumstances or genetic variation in the plants. Alternatively, the toxic principle may be an extrinsic substance that is not always present on or in grapes. There may also be large differences among individual dogs in their response to an intrinsic or an extrinsic component in the grapes. If such a difference among dogs exists and is genetically determined, then the response of the affected dogs could be considered idiosyncratic. 36 Spayed female dogs in this study were somewhat more likely to survive than were neutered male dogs, but the relevance of this finding is uncertain. Human males with chronic renal failure may be more likely to experience soft tissue calcification, with sex and peak Ca P being jointly associated in one retrospective study. 19 Sex-related differences in adverse reactions to xenobiotics in humans 37 and other species 38,39 also have been identified. Inclusion of more sexually intact dogs in the study would have improved the strength of the evaluation for sex-related differences. Determination of the mechanism of renal injury from grapes and raisins requires further evaluation and is beyond the scope of this study. The sequence of events seen in the dogs of this study is different from that reported for cholecalciferol rodenticide toxicosis, in which serum phosphorus and calcium concentrations and Ca P increase before serum creatinine and BUN concentrations due to the effects of hypercalcemia and mineralization on the kidneys. 40 Also, serum calcium and phosphorus concentrations are consistently increased in dogs with vitamin D toxicosis, whereas hypercalcemia and hyperphosphatemia were detected in 62.5% and 90% of the dogs in the present study, respectively. Additionally, evaluation of serum 25-(OH)D3 concentration in 1 dog failed to detect this persistent vitamin D metabolite. Although evaluation of 1 metabolite of vitamin D in 1 dog does not eliminate a vitamin D like effect from consideration as a contributing factor to ARF in these dogs, other features of

9 Grapes in Dogs 671 these cases also are dissimilar to those of cholecalciferol toxicosis, making a vitamin D like toxicosis unlikely. Findings in the dogs of this report also are different than those seen after ethylene glycol ingestion, in which azotemia often does not develop until 48 hours after exposure 41 and in which approximately half of the dogs become hypocalcemic. 31 There were no historical, laboratory, or histopathological findings indicative of ethylene glycol exposure in the dogs of the present study. Renal failure in cats due to lily ingestion (Lilium spp. and Hemerocallis spp.) has a similar rapid onset of signs when compared with the signs exhibited by these dogs. Generally, hypercalcemia does not develop in cats with ARF due to lily toxicosis, but hyperphosphatemia and severe azotemia typically are identified. 42 Renal tubular damage due to lily ingestion has not been seen in dogs, but pancreatitis (confirmed by histopathology) developed in 2 cats in ARF as a result of lily ingestion. 43 It is unknown whether or not the mechanism by which pancreatitis developed in one or more dogs in this study is similar to that in the 2 cats reported by Langston. Leptospirosis can cause ARF, with a clinical picture similar to that exhibited by the dogs in this study, 44 but hypercalcemia does not occur often in dogs with leptospirosis. 45 Three dogs in this study had detectable titers to various serovars of Leptospira, but none of the titers were high enough to be considered diagnostic for infection. 46 The dogs in this series did not exhibit clinical or clinicopathological signs consistent with acute bacterial pyelonephritis. 47 Additionally, evidence of pyelonephritis was not present in any of the histopathological samples of kidney that were evaluated. Mycotoxins produced by molds on grapes bear consideration as a possible cause for the signs in the dogs of this study. Ochratoxin A and citrinin, which are both produced by various Aspergillus and Penicillium species found on grapes and raisins, 48 have been shown experimentally to damage the proximal tubular epithelium in dogs. 49,50 Naturally occurring renal damage from ochratoxin A or citrinin generally is a result of chronic exposure and has a progressive clinical course. 51 Neither mycotoxin has been demonstrated to cause acute disease in dogs through environmental exposure. This may be because the amounts of these contaminants in fruit 48 are low enough that a dog would be unlikely to ingest enough fruit to reach the acute dosages revealed to cause lesions experimentally. The lack of ochratoxin A in 3 samples of grapes and raisins and the absence of other mycotoxins in a sample of raisins further reinforces that it is unlikely that the renal tubular necrosis seen in these dogs was caused by mycotoxicosis. The presence of pesticide residues on grapes and raisins also should be considered. Various fungicides, insecticides, and herbicides often are applied to Vitis spp. to prevent damage to the vines and the grapes from deleterious organisms. The concentrations of pesticides on the fruits often decline rapidly due to evaporation and degradation, and current evidence indicates that pesticide residues consistently are either not present or are present at below regulatory limit concentrations if the pesticides are applied according to good agricultural practice. 52 Reviewing information pertaining to grapes and raisins from the Food and Drug Administration s pesticide residue monitoring program reinforces this view. 53 In 2002, 100 domestic and imported samples of grapes and raisins were individually evaluated for residues of 266 pesticides. Although 2 imported samples of grapes were in violation, the 67 samples from Chile and Mexico (the primary exporters of fresh grapes to the United States) and all domestic samples were in compliance, revealing either no residues or residues that were measured to be below regulatory limits. Also, the findings that no chlorinated hydrocarbon, carbamate, and organophosphorus insecticides were found in a sample of raisins and that no herbicides were found in hepatic tissue from one dog in this study are encouraging, but a fungicide screen was not performed on either sample. Any pesticide residues present on grapes or raisins are unlikely to contain pesticide concentrations high enough to result in ARF. Given the potential harm that may occur after ingestion of grapes or raisins by dogs, performing basic decontamination should be considered if a dog is suspected to have ingested grapes or raisins. Rapid decontamination may be critical to prevent potential harm, and should be considered even if intervention is delayed. Raisins are hygroscopic and increase in volume as they absorb water in the gastric lumen. 54 The digestion of raisins and grapes into smaller particles in the stomach tends to be slow, as was evidenced in some of these dogs by the vomiting of swollen raisins or grapes after they had remained in the stomach overnight. Inducing emesis with standard emetics 55 may recover grapes or raisins several hours after exposure. Administration of activated charcoal 55 should be considered, especially if ingested grapes or raisins are not recovered by inducing emesis. Without a better understanding of the etiology of ARF in the dogs of this report, it is not certain whether or not activated charcoal is of benefit. The potential benefit of charcoal administration, however, likely outweighs the risks, but individual patient factors should be considered. Initiating fluid diuresis should be considered, despite the fact that the benefit of fluid therapy was uncertain in this study. Potential benefits of fluid diuresis include preservation of renal hemodynamics, increasing endogenous solute excretion, decreasing tubular obstruction due to necrosis, and potentially decreasing tubular reabsorption of a nephrotoxic agent by increasing tubular flow. 32 Early detection of the onset of ARF would be enhanced by monitoring for clinical signs, monitoring laboratory variables such as BUN concentration, serum creatinine, serum total calcium, and serum phosphorus concentrations, urine sediment findings and urine glucose, and Ca P. If clinical signs and azotemia in a dog are suspected to be due to the ingestion of grapes or raisins, it is important to determine whether or not any other causes or complicating factors could be contributing to the patient s condition. In this study, the results of radiography and ultrasonography were not consistently abnormal, and these imaging techniques are likely not to be useful in confirming that ARF is a result of ingestion of grapes or raisins. Thoracic and abdominal radiographs and abdominal ultrasonography, however, may offer other benefits, such as ruling out other nephrotoxicants (eg, ethylene glycol) 56 and identifying complications from ARF or its therapy, such as overhydration, pancreatitis, and pulmonary mineralization, all of