This work was partly delivered at the 25th ACVIM Forum, Seattle, WA, June 6–9, 2007 and 17th ECVIM-CA Congress, Budapest, Hungary, September 13–15, 2007.
Evaluation of a Portable Meter to Measure Ketonemia and Comparison with Ketonuria for the Diagnosis of Canine Diabetic Ketoacidosis
Version of Record online: 13 APR 2009
Copyright © 2009 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 23, Issue 3, pages 466–471, May/June 2009
How to Cite
Tommaso, M. D., Aste, G., Rocconi, F., Guglielmini, C. and Boari, A. (2009), Evaluation of a Portable Meter to Measure Ketonemia and Comparison with Ketonuria for the Diagnosis of Canine Diabetic Ketoacidosis. Journal of Veterinary Internal Medicine, 23: 466–471. doi: 10.1111/j.1939-1676.2009.0302.x
- Issue online: 19 MAY 2009
- Version of Record online: 13 APR 2009
- Submitted April 30, 2008; Revised February 12, 2009; Accepted February 12, 2009.
Background: The diagnosis of canine diabetic ketoacidosis (DKA) usually is based on measurement of urinary acetoacetate (ketonuria). In humans, this test is less sensitive and specific than blood 3-β-hydroxybutyrate (ketonemia) evaluation.
Hypothesis: Ketonemia measurement using a portable meter is more accurate than ketonuria determination with a dipstick to diagnose canine DKA.
Animals: Seventy-two client-owned diabetic dogs with ketonemia, ketonuria, or both.
Methods: Prospective observational study. Based on blood bicarbonate concentration and anion gap, dogs were divided into 2 groups: patients with DKA (n= 25); patients with diabetic ketosis (n= 47). Sensitivity, specificity, and positive and negative likelihood ratio (LR) at different cut-off points were determined for both ketonemia and ketonuria. Receiver operating characteristic (ROC) analysis was used to assess the accuracy of each diagnostic test to diagnose DKA.
Results: With regard to ketonemia, cut-off values of 2.3 and 4.3 mmol/L revealed 100% sensitivity and 100% specificity, respectively, whereas cut-off values of 2.8 and 3.5 mmol/L showed a −LR of 0.05 and a + LR of 13.16, respectively. With regard to ketonuria, a cut-off value of 1+ revealed 92% sensitivity, 40% specificity, and −LR of 0.20, whereas a cut-off value of 3+ revealed 44% sensitivity, 94% specificity, and +LR of 6.89. The areas under the ROC curves for the ketonemia and ketonuria tests were significantly different (0.97 and 0.81, respectively, P= .003).
Conclusions and Clinical Importance: Measurement of ketonemia is accurate and more effective than measurement of ketonuria to diagnose canine DKA.
partial pressure of CO2
receiver operating characteristic
Diabetic ketoacidosis (DKA) is an acute, life-threatening complication of diabetes mellitus that responds to prompt diagnosis and treatment both in humans and animals.1,2 DKA is characterized by the biochemical triad of hyperglycemia, hyperketonemia, and metabolic acidosis associated with severe metabolic deterioration.3 In DKA, absolute or relative insulin deficiency and simultaneous increases in circulating counterregulatory hormones, especially glucagon, caused by stressful events (eg, concurrent diseases) or drugs, combine to enhance lipolysis in adipose tissue and ketogenesis in the liver.4–6 Acetoacetate (AcAc) and 3-β-hydroxybutyrate (3HB), the most important ketone bodies produced during DKA, are strong organic acids that fully dissociate at physiological pH resulting in metabolic acidosis.7
The diagnosis of DKA in dogs is sometimes difficult and usually based on clinical signs and measurements of blood glucose, ketone bodies, electrolytes, and blood gases. Evaluation of acid-base status is used to determine the presence and severity of metabolic acidosis, whereas ketosis is evaluated by ketone body assessment.8,9 In humans, estimation of urinary or blood ketone bodies by means of commercial semiquantitative ketone tests (nitroprusside-reactive dipstick), which only measure AcAc, has been the traditional method to diagnose hyperketonemia. However, 3HB is the predominant ketone body in human DKA.10 Based on the severity of DKA, the serum 3HB/AcAc concentration ratio frequently increases to ≥ 3 : 1 as a result of the highly reduced state in hepatic mitochondria.11,12 Thus, severe hyperketonemia may be underestimated or undetected if only measurement of urinary or blood AcAc is used.13,14 Moreover, use of the nitroprusside-reactive dipstick is associated with considerable risk of false-positive and -negative results when some drugs are present in the urine or the dipstick has been exposed to air for a long time.15–17 Finally, obtaining a urine sample for a rapid diagnosis may be difficult in emergency cases, and the visual diagnosis of ketone bodies using the dipstick is qualitative and subjective, and subject to interpretive errors.18 For these reasons, blood 3HB testing methods are now preferred over urinary or blood AcAc testing to diagnose and monitor DKA in humans.10
Estimation of urinary AcAc is the preferred method to diagnose canine DKA.3 To the authors' knowledge, few reports have described the use of 3HB evaluation in DKA-affected dogs.6,19 One of these studies showed that serum 3HB, expressed as a percentage of serum ketone concentration, decreased as total serum ketone increased and, therefore, 3HB might not be the predominant ketone body in DKA-affected dogs.6
In recent years, a hand-held electrochemical sensor that measures venous or capillary blood 3HB concentration in humans has become commercially available.20–24
The aim of the present study was to compare the diagnostic accuracy of blood 3HB measurement by means of an electrochemical sensor with that of urinary AcAc determination for the diagnosis of DKA in the dog.
Materials and Methods
Criteria for the Selection of Cases
A convenience sample of diabetic dogs (newly diagnosed and insulin-treated), admitted to the Department of Veterinary Clinical Sciences of the University of Teramo between January 2004 and January 2007 with clinical and laboratory signs of ketosis or ketoacidosis, was included in this prospective observational study. Seventy-two dogs met the following inclusion criteria: blood glucose >250 mg/dL, glucosuria and ketonuria (urinary AcAc ≥1+), ketonemia (blood 3HB ≥0.1 mmol/L), or both. Based on blood bicarbonate (HCO3−) concentration and anion gap (AG), dogs were divided into 2 groups: patients with DKA (DKA group) with HCO3− <15 mEq/L and AG ≥ 20 mEq/L and patients with diabetic ketosis (DK group) with HCO3−≥ 15 mEq/L and AG <20 mEq/L. Before inclusion of dogs in the study, informed consent was obtained from owners and the protocol was approved by the Ethics Committee of the University of Teramo.
Sample Collection and Analytical Procedures
After physical examination, blood samples from each dog were collected by jugular venipuncture for routine CBC, biochemical profile, and serum 3HB concentrations. Furthermore, direct evaluation of glycemia and ketonemia was carried out using a glucometera and hand-held electrochemical ketone sensor,b respectively. According to the manufacturer's recommendations, an electrochemical strip was inserted into the sensor to which 10 μL of whole venous blood were applied and after 30 seconds, the blood 3HB concentrations were displayed. The assay range is 0.0–6.0 mmol/L. Because the blood ketone meter displays “HI” (for high) when blood ketones are >6 mmol/L, we considered the “HI” value to be equal to 6 mmol/L. The analytical accuracy of the employed portable ketone meter was evaluated in our laboratory by comparing its measurements of blood 3HB to those determined on serum by a standard laboratory enzymatic methodc adapted for an automatic analyzer.d Regression and Bland-Altman analyses showed a good agreement between the 2 methods (Fig 1). Within 10 minutes, urine specimens for urinalysis and heparinized venous blood samples for blood gas and electrolyte analysis were collected by means of cystocentesis and jugular venipuncture, respectively, and immediately processed (within 15 minutes and 2 minutes after the collection, respectively). Urine samples were analyzed by use of 10-patch test stripse that include semiquantitative estimation of ketonuria with an assay range from negative (0) to 150 mg/dL (3+).
Venous blood pH and partial pressure of CO2 (PCO2) were measured with a blood gas analyzer.f Electrolyte concentrations were determined in whole blood by the same analyzer with ion-specific electrodes. The HCO3− concentrations were calculated automatically by the instrument based on measurement of pH and PCO2, and AG was calculated by subtracting (HCO3−+ Cl−) from (Na++ K+).
Data analysis was performed by statistical software packages.g,h In order to compare the distribution of data between the groups, the t-test for normally distributed continuous data, the χ2 test for categorical data, and the Mann-Whitney U-test for ordinal data, were used. The sensitivity, specificity, and negative and positive likelihood ratios (−LR and +LR, respectively) were calculated for each of the 2 tests (ie, ketonemia and ketonuria) at different cut-off points. Sensitivity and specificity values were used to generate receiver operating characteristic (ROC) curves in order to assess the overall accuracy of both diagnostic methods, determined by calculating the area under the ROC curve. Furthermore, the accuracy of the 2 tests was compared by comparing their ROC curves. For all statistical analyses, a P value < .05 was considered significant.
Patients' Characteristics and Laboratory Findings
Forty dogs were newly diagnosed diabetics and 32 were insulin-treated diabetics. There were no statistical differences between the 2 groups on the basis of age (range, 3–15 years; mean, 8.5 years) and body weight (range, 2–36 kg; mean, 13.7 kg). There was a predominance of mixed breed (n= 43) and female (n= 57) dogs in both groups. Twenty-seven females were intact and 30 were spayed, whereas 4 of the 15 male dogs were neutered. The DKA group consisted of 25 dogs, whereas the DK group was composed of 47 dogs. Concurrent disorders or precipitant factors for DKA dogs were pancreatitis (n= 4), omission of or inadequate insulin therapy (n= 4), corticosteroid therapy (n= 3), diestrus (n= 3), pyometra (n= 3), and 1 each for hyperadrenocorticism, urinary tract infection, pheochromocytoma, renal failure, mammary carcinoma, and pyelonephritis. In 2 dogs with DKA, there was no evident precipitant condition. Prevalent concurrent diseases in DK dogs were hyperadrenocorticism (n= 9) and urinary tract infection (n= 7).
Results of blood glucose concentrations, venous blood HCO3−, AG, ketonemia, and ketonuria are shown in Table 1. Blood glucose concentrations were significantly higher in DKA compared with DK dogs (P < .01) as were ketonemia and ketonuria results (P < .001).
|DK (n= 47)||DKA (n= 25)|
|Mean ± SD||No. (%)||Mean ± SD||No. (%)|
|Glycemia (mg/dL)abc||390 ± 105||462 ± 107|
|Venous blood HCO3− (mEq/L)d||22.8 ± 3.7||10.7 ± 3.4|
|AG (mEq/L)d||16.9 ± 3.6||27.2 ± 3.8|
|Blood 3HB (mmol/L)ebf||1.5 ± 1.2||4.7 ± 1.2|
|0 (negative)||19 (40.4)||2 (8)|
|1 + (10 mg/dL)||13 (27.7)||2 (8)|
|2 + (50 mg/dL)||12 (25.5)||10 (40)|
|3 + (150 mg/dL)||3 (6.4)||11 (44)|
Performance of Ketonemia and Ketonuria Tests
The diagnostic efficacy of ketonemia and ketonuria tests to detect DKA is shown in Appendix 1. In particular, the cut-off value of 2.3 mmol/L for ketonemia showed the best sensitivity (100%) with 70% specificity, whereas the cut-off value of 4.3 mmol/L showed the best specificity (100%) with 64% sensitivity to diagnose DKA. The cut-off value of 3.5 mmol/L for ketonemia showed a +LR of 13.16, whereas the cut-off value of 2.8 mmol/L showed a −LR of 0.05.
The cut-off value of 0 of ketonuria showed 100% sensitivity and 0% specificity, whereas the cut-off value of 3+ showed 94% specificity and 44% sensitivity. The cut-off value of 3+ for ketonuria showed a +LR of 6.89, whereas the cut-off value of 1+ showed a −LR of 0.20.
The overall accuracy of ketonemia and ketonuria tests was 0.97 (95% confidence interval [CI], 0.89–0.99) and 0.81 (95% CI, 0.70–0.89), respectively (Fig 2). Comparison of the ROC curves for the 2 tests showed a significant difference between the areas under the ROC curve (0.16; 95% CI, 0.06–0.26; P= .003).
Venous blood 3HB concentration was significantly higher in DKA compared with DK dogs in the present study, as were urinary AcAc concentrations. Ketone concentrations ranging from 2.3 (sensitivity 100%) to 4.3 mmol/L (specificity 100%) showed wide variability between sensitivity and specificity and no cut-off value was able to discriminate the 2 stages of the disease, and ketonuria values that ranged from negative (sensitivity 100%) to 3+ (specificity 94%). However, the diagnostic accuracy of ketonemia measurement to predict DKA was very high (area under the ROC curve =0.97), whereas ketonuria determination was moderately accurate (area under the ROC curve =0.81) and significantly lower. Furthermore, ketone concentrations ≥3.5 mmol/L predict DKA with a +LR > 13, whereas a ketone concentration <2.8 mmol/L excludes DKA with a −LR of 0.05 in dogs of this study. On the contrary, low +LR (6.89 for 3+ of AcAc reading) and high −LR (0.2 for 1+) were obtained using a commercial urine dipstick.
Based on the results of this study, specific management of DKA should be initiated in dogs with ketonemia ≥3.5 mmol/L, because they are at higher risk of presenting with overt clinical DKA, similar to human patients.25 On the contrary, when the ketone concentration is <2.8 mmol/L the risk of DKA is very low. Data obtained using the urine dipstick produced moderate changes in posttest probability estimates,26 therefore they did not allow conclusive decisions for the diagnosis of DKA in dogs. Use of the urinary dipstick has been associated with a substantial risk of both false-positive and -negative results in human patients. In the present study, ketonuria was either undetectable or equal to 1+ in 4 dogs with severe DKA and, conversely, equal to 3+ in 3 dogs with DK. Thus, absence of ketonuria did not exclude DKA in dogs and, conversely, presence of ketone bodies in the urine may lead to false-positive diagnoses in dogs without DKA. These findings may be caused by a predominance of 3HB over AcAc in DKA-affected dogs, although 3HB was not the prevalent ketone body in another study.6
The results for ketonemia in this study are similar to those reported previously in 43 and 41 dogs with DKA and DK, respectively.19 The observed range of ketonemia (2.3–4.3 mmol/L), with overlapping of DKA and DK cases, may be because of a different trend in serum 3HB/AcAc ratio in diabetic dogs compared with humans, thus suggesting that 3HB may not be the predominant ketone body in DKA-affected dogs.
No standard laboratory criteria exist to define DKA in animals19,27–31 and this lack of standardization may further explain the overlapping of DKA and DK cases. According to the guidelines of the American Diabetes Association,32 also applied to dogs and cats,19,28 a HCO3− concentration ≤15 mEq/L was used as a cut-off in dogs of the present study. For better identification of subjects with metabolic acidosis, only those dogs with concurrently increased AG, which results from accumulation of ketoacids,7,9 were included in the DKA group, as has been done in other studies.31,33,34 Nevertheless, our data confirm the poor performance of an increased AG for the diagnosis of DKA in dogs.35 The overlapping of dogs with DKA and DK observed in the present study also may be attributable to augmented urinary excretion of ketoanions in dogs with DKA but without severe volume depletion, thus preventing an increased AG metabolic acidosis.7 Another cause of increased AG may be increased serum concentrations of other unmeasured anions (eg, lactate, phosphate, and sulfate) leading to acidosis in DKA patients,7 although acidosis in most diabetic dogs is primarily attributable to hyperketonemia.6 The lack of estimation of the effects of serum lactate concentration and possible renal failure on increased AG metabolic acidosis is a limitation of the present study. Furthermore, we calculated AG by subtracting the major measured anions (Cl− and HCO3−) from the major measured cations (Na+ and K+). However, because K+ concentrations may be altered by acid-base disturbances and total body stores, it is not routinely used for the calculation of AG in humans.36
In conclusion, the handheld electrochemical ketone sensor measuring blood 3HB was shown to be more accurate than urinary AcAc tests for the identification of DKA-affected dogs. Furthermore, considering that DKA is a true medical emergency requiring prompt therapeutic management, its easy usage, the small volume sample required, and the short time to obtain test results make it a very useful device for the identification of DKA. As already demonstrated in humans, accepted guidelines for the interpretation of ketone concentrations are necessary in order to clearly identify DKA-affected dogs.
a Accu-Chek Comfort, Roche Diagnostics GmbH, Mannheim, Germany
b MediSense Optium, Abbott Laboratories, Oxon, UK
c Procedure No. 310-UV, Sigma Diagnostics, St Louis, MO
d AU400, Mishima Olympus Co Ltd, Shizuoka, Japan
e Combur-test10 UX, Roche Diagnostics GmbH
f Synthesis 20, Instrumentation Laboratory Spa, Milan, Italy
g SPSS 13.0 for Windows, SPSS Inc, Chicago, IL
h MedCalc 7.3, MedCalc Software, Mariakerke, Belgium
- 1Diabetic ketoacidosis: Treatment recommendations. Comp Cont Ed Pract Vet 2001;23:330–340.
- 3Diabetic ketoacidosis. In: FeldmanEC, NelsonRW, eds. Canine and Feline Endocrinology and Reproduction, 3rd ed. St. Louis, MO: Saunders; 2004:580–615.,
- 7Metabolic acid-base disorders. In: DiBartolaSP, ed. Fluid, Electrolyte, and Acid-Base Disorders in Small Animal Practice, 3rd ed. St. Louis, MO: Saunders Elsevier; 2006:251–283.
- 8Diabetic ketoacidosis: Pathophysiology and clinical laboratory presentation. Comp Cont Ed Pract Vet 2001;23:220–228.
- 10American Diabetes Association. Test of glycemia in diabetes. Diabetes Care 2004;27 (Suppl 1):S91–S93.
- 17Contamination of a specimen with N-acetyl cysteine infusion: A cause of spurious ketonaemia and hyperglycaemia. Ann Clin Biochem 1989;26:207., ,
- 22Accuracy of an electrochemical sensor for measuring capillary blood ketones by fingerstick samples during metabolic deterioration after continuous subcutaneous insulin infusion interruption in type 1 diabetic patients. Diabetes Care 2003;26:1137–1141., , , et al.
- 30Outcome of dogs with diabetic ketoacidosis: 127 dogs (1993–2003). J Vet Intern Med 2006;20:547–555., ,
- 31Plasma acetoacetate measurements using urinary dip stick methodology as a screen for ketonaemia in feline diabetic ketoacidosis. Proceeding 17th ECVIM-CA Congress Budapest, Hungary, 2007:212.,
- 32American Diabetes Association. Hospital admission guidelines for diabetes mellitus. Diabetes Care 2001;24 (Suppl 1):S91.
- 35Performance of various indexes as screening test for marked ketonemia in dogs with diabetes mellitus. Proceeding 25th ACVIM Forum Seattle, WA, 2007:842–843., , ,
- 36Diabetic ketoacidosis. A combined metabolic-nephrologic approach to therapy. Diabetes Rev 1994;2:209–238., ,
|%||95% CI||%||95% CI|