The study was done at the Clinic for Small Animal Internal Medicine and Institute of Veterinary Physiology, Vetsuisse Faculty, University of Zurich, Zurich, Switzerland.
Remission of Diabetes Mellitus in Cats Cannot be Predicted by the Arginine Stimulation Test
Article first published online: 10 DEC 2010
Copyright © 2010 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 25, Issue 1, pages 83–89, January/February 2011
How to Cite
Tschuor, F., Zini, E., Schellenberg, S., Wenger, M., Kaufmann, K., Furrer, D., Lutz, T.A. and Reusch, C.E. (2011), Remission of Diabetes Mellitus in Cats Cannot be Predicted by the Arginine Stimulation Test. Journal of Veterinary Internal Medicine, 25: 83–89. doi: 10.1111/j.1939-1676.2010.0649.x
- Issue published online: 11 JAN 2011
- Article first published online: 10 DEC 2010
- Submitted April 15, 2010; Revised September 30, 2010; Accepted October 15, 2010.
Background: Cats with diabetes mellitus frequently achieve clinical remission, suggesting residual β-cell function. Responsiveness of β-cells to arginine persists the longest during diabetes progression, making the intravenous arginine stimulation test (IVAST) a useful tool to assess residual insulin and glucagon secretion.
Hypothesis: Diabetic cats with and without remission will have different arginine-induced insulin or glucagon response.
Animals: Seventeen cats with diabetes, 7 healthy cats.
Methods: Blood samples collected on admission and during subsequent IVAST. Glucose, insulin, and glucagon were measured. Response to IVAST was assessed by calculating the insulin and glucagon area under the curve (AUC) and the AUC glucagon-to-insulin ratio. Diabetic cats were treated with insulin and were followed for 18 weeks. Remission was defined as normoglycemia and disappearance of clinical signs of diabetes for ≥4 weeks, without requiring insulin.
Results: Seven diabetic cats (41%) achieved remission. On admission, blood glucose concentration was significantly lower in cats with remission (median, 389 mg/dL; range, 342–536 mg/dL) than in those without remission (median, 506 mg/dL; range, 266–738 mg/dL). After IVAST, diabetic cats with remission had higher AUC glucagon-to-insulin ratios (median, 61; range, 34–852) than did cats without remission (median, 26; range, 20–498); glucose, insulin, and glucagon AUCs were not different. Diabetic cats had lower insulin AUC than did healthy cats but comparable glucagon AUC.
Conclusions and Clinical Importance: Diabetic cats with and without remission have similar arginine-stimulated insulin secretion on admission. Although cats with remission had lower blood glucose concentrations and higher AUC glucagon-to-insulin ratios, large overlap between groups prevents use of these parameters in clinical practice.
area under the curve
- G0, I0, and Gl0
glucose, insulin, and glucagon concentrations before arginine stimulation, respectively
- IPR and GlPR
insulin and glucagon peak response, respectively
intravenous arginine stimulation test
Diabetes mellitus (DM) is one of the most common endocrinopathies in cats, and its incidence is increasing because of a rise in predisposing factors, such as obesity and physical inactivity.1,2 Most cats seem to suffer from a Type 2-like form of DM that is characterized by decreased insulin secretion and insulin resistance. In up to 50% of diabetic cats, insulin therapy can be withdrawn within 4 months after beginning treatment.3–6 This phenomenon, called diabetic remission, is thought to be because of recovery from glucose toxicity.7
Several insulin secretagogue tests have been evaluated for measurement of insulin secretion capacity in humans, including the hyperglycaemic clamp, the intravenous (IV) and by mouth (per os, PO) glucose tolerance tests, the IV glucagon stimulation test, and the intravenous arginine stimulation test (IVAST).8–12 In Type 1 or 2 diabetic humans, β-cells show progressive deterioration in their responsiveness to various secretagogues (eg, glucose, glucagon, amino acids).13,14 Responsiveness to amino acids has been shown to outlast other stimuli, thus suggesting that the IVAST may be used to detect residual β-cell secretory capacity later during the progression of DM.15l-Arginine is known to be the most potent insulin secretagogue of all amino acids. It increases β-cell secretion by membrane depolarization and a subsequent increase in intracellular calcium.16 Apart from β-cells, l-arginine also stimulates glucagon from pancreatic α-cells by a similar mechanism.12
Only 1 study has investigated residual insulin secretory capacity in cats with DM,4 but the glucagon stimulation test failed to show significant differences in insulin response between cats that achieved diabetic remission and those that did not.4
The IVAST so far has been investigated in healthy cats and was shown to be a valuable tool for evaluating insulin secretory capacity.a,17 The test also was used for simultaneous stimulation of α-cells, thus yielding information about glucagon secretion. Until now, the IVAST has not been used in diabetic cats, and it is therefore unknown whether it can help differentiate between cats that achieve diabetic remission and those that do not during the course of treatment. Therefore, the objectives of this study were to assess insulin and glucagon response in an IVAST in healthy and diabetic cats and to evaluate if differences in α- and β-cell response to arginine exist in cats with DM that will achieve remission as compared with those that will not.
Materials and Methods
Seven healthy control and 17 diabetic cats were included in the study. Seven neutered male, healthy domestic shorthair catsb were used following principles of laboratory animal care (permission No. 213/2003, Veterinary Office of Zurich, Switzerland). The median age of control cats was 15 months (range, 16–24 months), and they were healthy based on physical examination and routine clinical clinicopathologic data; cats had a body condition score of 5 (ie, lean cats) and their median body weight was 4.8 kg (range, 4.5–5.2 kg). Cats had free access to water and were fed twice daily with a commercial dietc given at maintenance. The diet was given for 4 consecutive weeks before testing.
Cats admitted to the Clinic for Small Animal Internal Medicine, Vetsuisse Faculty, University of Zurich between September 2004 and May 2008 were considered for the study if DM was diagnosed. A diagnosis of DM was based on clinical signs (eg, polyuria, polydipsia, weight loss), hyperglycemia (fasting blood glucose >180 mg/dL), glucosuria, and increased serum fructosamine concentration (>340 μmol/L; reference range, 200–340 μmol/L). Diabetic cats were included if they had not received insulin before admission and if diabetic ketoacidosis, acromegaly, heart failure, kidney failure, urinary tract infection, and hyperthyroidism were absent based on clinical findings, plasma IGF-1 measurement, serum biochemistry, urinalysis, urine culture, and diagnostic imaging. Cats that had received corticosteroids or progestagens within 6 months before admission and those with suspected pancreatitis based on clinical findings and abdominal ultrasound examination also were excluded. To be included in the study, all cats must have had follow-up of at least 18 weeks.
In healthy cats, a central venous catheter was implanted in the jugular vein on the day before IVAST. The procedure was carried out under general anesthesia. In diabetic cats, after routine diagnostic evaluation, a central venous catheterd was placed in the saphenous vein under general anesthesia.
The IVAST was performed on the day after implantation of the central venous catheter in cats that had been fasted overnight and not treated with insulin. Diabetic cats received IV fluids (0.9% NaCl at 1 mL/kg/h supplemented with 20 mEq of KCl/L of NaCl) to minimize the risk of development of ketoacidosis. A baseline blood sample was collected and thereafter argininee was infused slowly (over 1 minute) via a cephalic vein at a dosage of 0.2 g/kg body weight.17 Blood samples were obtained after 2, 4, 7, 9, 15, 25, and 30 minutes via the central venous catheter. Baseline samples and samples collected after arginine injection were immediately placed in ice-cold EDTA tubes containing aprotininef and centrifuged within 5 minutes after the test. Plasma was stored at −80°C until further use. At baseline, glucose (G0), fructosamine, insulin (I0), and glucagon (Gl0) were measured, and the Gl0-to-I0 ratio was calculated. Insulin and glucagon responses to IVAST were assessed by determining the insulin and glucagon peak response (IPR and GlPR, respectively) above baseline by calculating the insulin and glucagon area under the curve (AUC) during the initial 9 minutes (AUC9) and the entire 30 minutes (AUC30) of stimulation, and by calculating the AUC9 and AUC30 glucagon-to-insulin ratio. The reason for calculating the AUC9, in addition to AUC30, is that the highest response to arginine has been demonstrated during the 1st 9 minutes after injection in cats.17 Insulin-to-glucagon ratios were calculated because it has been shown that a close interplay exists between α- and β-cells in mice and rats, with glucagon being 1 possible α-cell product necessary for normal β-cell secretion.18
Treatment of Cats with DM
After recovery from the IVAST, diabetic cats were discharged and prescribed routine SC treatment with insulin, either porcine insulin zinc suspensiong or insulin glargine,h based on ease of availability, and according to the following scheme: the starting dose in cats ≤4 kg was 1 U q12h and in cats >4 kg 1.5–2 U q12h. In addition, a specific commercial dietc was recommended. Follow-up examinations were performed 1, 3, 6, 10, and 18 weeks after admission and included assessment of clinical signs and body weight, measurement of a serum biochemical profile with serum fructosamine concentration, and generation of a blood glucose curve. The latter was performed by measuring capillary blood glucose every 2 hours over 10–12 hours, as previously described.19 The insulin dose was adjusted to improve glycemic control if necessary. Because the ideal glucose nadir should fall between 90 and 160 mg/dL, if the nadir was below 90 mg/dL the insulin dose was decreased by 0.5–1 U per injection; if the nadir was above 160 mg/dL, the insulin dose was increased by 0.5–1 U per injection. The diet was changed if the cat developed renal failure, diarrhea, or food aversion.
Cats with and without Remission of DM
To achieve remission, cats needed to have no clinical signs of diabetes (eg, polyphagia, polyuria, polydipsia), as well as normoglycemia and normal serum fructosamine concentrations for at least 4 weeks, without insulin administration.20 Cats that required insulin throughout the study were defined as not being in remission.
CBCs, serum biochemical profiles, and urinalyses were performed by standard laboratory methods. Plasma glucose and serum fructosamine concentrations were measured by an automatic analyzeri with commercial reagents.j Insulin and glucagon were measured by commercial radioimmunoassayk,l previously validated in our laboratory.21,22 Insulin interassay and intra-assay coefficients of variation were 7.0 and 6.5%, respectively; the sensitivity of the assay was 2 μU/mL. To validate the glucagon assay, plasma from 5 healthy cats was assayed 7 times with 2 different kits. To determine sensitivity, the 95% probability of the zero standard and the lowest standard concentration that was significantly different from zero were measured, and the average value was calculated. Parallelism was determined measuring a sample with a concentration of 510 pg/mL, serially diluted (ie, 100, 80, 60, 40, and 20%). Glucagon interassay and intra-assay coefficients of variation were 8.2 and 7.5%, respectively; the sensitivity of the assay was 50 pg/mL. With the above dilutions assay linearity was 108%.
Results are expressed as median and range. Characteristics of diabetic cats, including body weight, blood glucose, and fructosamine concentrations on admission, and prescribed dosage of insulin at discharge were compared between cats that achieved remission and those that did not by the Mann-Whitney U-test. The same test was used to compare results of the IVAST in healthy and diabetic cats and in diabetic cats with and without remission. Specifically, baseline concentrations of G0, I0, and Gl0, and the Gl0-to-I0 ratio, and, after arginine stimulation, the IPR, GlPR, AUC9, and AUC30 of glucose, insulin, and glucagon, and the AUC9 and AUC30 of glucagon-to-insulin ratios were determined.
The Wilcoxon matched paired test was used to assess whether initial body weight, blood glucose, and fructosamine concentrations at the time of admission differed between diabetic cats with and without remission after 18 weeks of treatment.
Results were considered significantly different at P < .05. Statistical analysis was performed by standard software.m
Cats with DM
Seventeen diabetic cats met the inclusion criteria. Median age was 12 years (range, 8–17 years). On admission, median body weight was 5.7 kg (range, 2.9–10 kg). Eleven cats were neutered males and 6 were spayed females; they were all domestic shorthairs. Median fasting blood glucose was 437 mg/dL (range, 266–738 mg/dL) and median fructosamine concentration was 666 μmol/L (range, 580–895 μmol/L). All cats had severe glucosuria; none had ketonuria or positive urine culture results.
Seven cats (41.2%) underwent remission during the study period, whereas remission did not occur in 10 cats (58.8%) (Table 1). Insulin in cats with remission could be discontinued after a median time of 8 weeks (range, 6–14 weeks). Age in cats with and without remission was similar. On admission, body weight was not different between groups and, within each group, did not change significantly after 18 weeks compared with baseline. Interestingly, on admission blood glucose concentration was significantly lower (approximately 25%) in cats that achieved remission, whereas fructosamine did not differ (Table 1). Blood glucose and fructosamine concentrations were significantly decreased in both groups after 18 weeks compared with admission. In all cats with remission, blood glucose and fructosamine concentrations were within normal limits after discontinuation of insulin therapy.
|Parameter||Unit||Remission (n = 7)||No Remission (n = 10)||P-Value|
IVAST in Healthy Cats and Cats with DM
Results of the IVAST in healthy and diabetic cats are presented in Table 2. Cats with DM had significantly higher G0 (P= .004), glucose AUC9 (P < .001), and glucose AUC30 (P < .001) than did healthy cats, whereas I0, IPR, insulin AUC9, and insulin AUC30 were significantly lower (P= .026, P < .001, P < .001, and P < .001, respectively). Gl0, GlPR, glucagon AUC9, and glucagon AUC30 were not different between healthy and diabetic cats. Of note, the glucagon AUC9 and glucagon AUC30 of 5 diabetic cats (4 with remission and 1 without remission) were 2 to 4 times higher than those of healthy cats. The Gl0-to-I0 ratio was not different between healthy and diabetic cats, whereas the AUC9 and AUC30 glucagon-to-insulin ratios were significantly higher in the latter group (P= .002 and .011, respectively). Insulin and glucagon responses to arginine in healthy cats and cats with DM are shown in Figure 1.
|Parameter||Unit||Healthy Cats (n = 7)||Diabetic Cats (n = 17)||P-Value|
|AUC9 glucose||mg/dL/9 min||666||558–828||2052||1026–3474||.001|
|AUC30 glucose||mg/dL/30 min||2196||1818–2628||8514||4176–13950||.001|
|AUC9 insulin||μU/mL/9 min||255||163–494||50||18–93||.001|
|AUC30 insulin||μU/mL/30 min||491||376–1320||196||88–309||.001|
|AUC9 glucagon||pg/mL/9 min||4331||2671–10190||3568||1055–52570||.899|
|AUC30 glucagon||pg/mL/30 min||13690||6752–33600||12670||3839–138200||.924|
|AUC9 glucagon-to-insulin ratio||15||8–63||74||25–1158||.002|
|AUC30 glucagon-to-insulin ratio||24||8–89||90||20–852||.011|
IVAST in Cats with and without Remission of DM
Results of the IVAST in diabetic cats that achieved remission and in those that did not are presented in Table 3. Differences were not observed between the 2 groups for G0 (ie, just before l-arginine administration), glucose AUC9 and AUC30, I0, IPR, insulin AUC9 and AUC30, Gl0, GlPR, glucagon AUC9, and the Gl0-to-I0 ratio. The AUC30 glucagon-to-insulin ratio was significantly higher in cats with remission (P= .033). Insulin and glucagon responses to arginine in diabetic cats with and without remission are shown in Figure 2.
|Parameter||Unit||Diabetic Cats with Remission (n = 7)||Diabetic cats without Remission (n = 10)||P-Value|
|AUC9 glucose||mg/dL/9 min||1944||1476–3006||2322||1026–3474||.417|
|AUC30 glucose||mg/dL/30 min||8226||6066–12078||9270||4176–13950||.417|
|AUC9 insulin||μU/ml/9 min||52||31–94||50||18–81||.740|
|AUC30 insulin||μU/mL/30 min||198||88–309||191||91–259||.887|
|AUC9 glucagon||pg/mL/9 min||19690||1170–52570||3365||1055–21330||.089|
|AUC30 glucagon||pg/mL/30 min||61500||6395–138200||10070||3839–92690||.055|
|AUC9 glucagon-to-insulin ratio||301||32–1158||70||25–429||.055|
|AUC30 glucagon-to-insulin ratio||61||34–852||26||20–498||.033|
In the present study, the insulin response to arginine was similar between cats with newly diagnosed DM that achieved remission to those that did not experience remission. In addition, we observed that baseline glucose concentrations, measured on admission, and the AUC30 glucagon-to-insulin ratio after arginine stimulation were higher in diabetic cats with remission, although a large overlap with cats that continued to require insulin existed. We also showed that diabetic cats, as compared with healthy cats, had decreased insulin but comparable glucagon secretion after IVAST.
Based on the notion that the insulin secretory response to arginine persists longer than the response to glucose and glucagon in diabetic humans,13–15 we hypothesized that the IVAST used at admission may identify cats with DM that could achieve remission. However, contrary to our hypothesis, insulin secretion after arginine was similar between both groups of diabetic cats, suggesting that the test is not sensitive enough to identify cats with residual β-cell function. Arginine-induced glucagon secretion also was comparable between both groups of diabetic cats, although it tended to be higher in cats with remission, as shown by the slightly increased AUC9 and AUC30 (P= .089 and .055, respectively). Notably, the AUC30 glucagon-to-insulin ratio was significantly higher (approximately 2.5-fold) in cats that achieved remission as compared with cats that continued to receive insulin. The reason diabetic cats with remission have an increased secretory reserve of glucagon relative to insulin is unclear, especially considering that the former hormone is diabetogenic and a potent stimulator of hepatic glycogenolysis and gluconeogenesis.23 However, pancreatic α-cells and glucagon seem to be required to maintain β-cell responsiveness to glucose.24 Indeed, glucagon receptor knock-out mice have impaired β-cell function.25 Furthermore, transgenic mice overexpressing the glucagon receptor in β-cells confirmed this concept, as the model showed improved glucose tolerance and increased insulin secretion in response to glucose.26 Whether the present findings in cats indicate a direct role for glucagon in supporting β-cells and insulin secretion remains unknown.
Compared with diabetic cats that required insulin during the study period, we found that the severity of hyperglycemia on 1st admission was lower in cats undergoing remission. This result suggests that blood glucose concentrations may be linked to the propensity for clinical recovery from DM in cats. Excessive hyperglycemia eventually may lead to severe and irreversible toxicity in feline β-cells,21 eventually preventing remission. However, similar concentrations of serum fructosamine were found in cats with or without remission, implying that the severity of hyperglycemia during the weeks before diagnosis was comparable in the 2 groups. Furthermore, another study performed at our institution showed that diabetic cats with and without remission had similar blood glucose concentrations at 1st diagnosis.6 The reason the less severe hyperglycemia in the cats of the present study that later did not need insulin remains unanswered. Of note, the difference in severity of hyperglycemia documented at admission waned after overnight saline infusion. Indeed, G0 (just before the IVAST) was not different between diabetic cats with and without remission. Saline infusion may have contributed to increased glucose excretion by the kidneys, thus decreasing the severity of hyperglycemia.
As expected, all diabetic cats had lower I0 than was observed in healthy controls. Arginine-induced insulin secretion also was decreased, as shown by lower IPR and insulin AUC9 and AUC30, suggesting β-cell dysfunction. Impaired insulin secretion may be because of β-cell exhaustion or decreased insulin gene expression.21 A severe decrease of islet cells is consistently observed in pancreatic tissue sections of diabetic cats,27 and β-cell loss also may have partly contributed to decreased insulin secretion.
Arginine is known to stimulate glucagon release in humans and rodents, increasing hepatic gluconeogenesis.28 This mechanism is believed to prevent hypoglycemia caused by amino acids absorbed through the gastrointestinal tract that would otherwise promote insulin secretion.29 Similar to a previous study we observed a 2-fold increase in glucagon concentrations 2 minutes after arginine administration in healthy cats (Fig 1B).17 In diabetic cats, glucagon secretion was not different compared with controls. Although recent studies demonstrated a positive role for glucagon in β-cells, it has been hypothesized that glucagon may be detrimental because it is associated with hyperglycemia in diabetic humans and in rodent models.30 Glucagon concentrations often are increased in various forms of DM in these species.31–33 Five diabetic cats had increased glucagon AUC9 and AUC30, suggesting that in some cases DM is associated with an increased glucagon response to arginine, as in humans and rodents. Because 4 of these 5 diabetic cats later achieved remission, an increased arginine-induced α-cell response may be associated with a favorable outcome in DM in cats.
Some limitations of this study should be mentioned. Diabetic cats were treated with either porcine insulin zinc suspensiong or insulin glargine,h possibly leading to bias in the results. In particular, as recently described, cats receiving insulin glargine may have increased likelihood of remission.34 However, in our series, only 2 of the 9 (22.2%) diabetic cats treated with insulin glargine achieved remission compared with 5 of the 8 (62.5%) diabetic cats that were treated with porcine insulin zinc suspension.g Thus, the potential beneficial effect of insulin glargine likely did not influence the frequency of remission in the present series.
Each of the diabetic cats was evaluated for 18 weeks after diagnosis. Some of the cats may have achieved remission after that period, if they had been followed-up for a longer time period. However, most diabetic cats achieving remission have insulin therapy discontinued within the 1st 4 months of treatment.3–6 Thus, the risk of having included diabetic cats with remission in the group without remission likely is low.
The group of diabetic cats was compared with healthy controls that were younger and male. Despite the fact that the group of diabetic cats included some females, male cats were chosen for the control group because diabetes is more commonly diagnosed in male cats. Diabetes is also generally diagnosed at an older age, thus it is possible that interpretation of IVAST results partially was biased by sex and age. Healthy cats were fed the same diet before IVAST, whereas diabetic cats received a variety of diets until diagnosis of diabetes was made and tests were performed. Thus, we cannot rule out an effect of diet on test results.
In summary, the similar arginine-stimulated insulin response in diabetic cats with and without remission suggests that the IVAST cannot be used to predict those cats with adequate residual β-cell function at diagnosis. The higher AUC30 glucagon-to-insulin ratio observed in diabetic cats that did not require insulin to maintain normoglycemia may indicate that a relative increase of α-cell function is involved in the mechanisms leading to remission. Less severe hyperglycemia on admission in cats undergoing remission has not been previously reported and warrants additional confirmation. Unfortunately, the large overlap between results of the AUC30 glucagon-to-insulin ratio and blood glucose concentrations prevents the use of these parameters to reliably predict diabetic cats with remission in clinical practice.
a Link KR, Rand JS. Arginine and phentolamine response test in cats. J Vet Intern Med 1996;146:185 (abstract)
b Harlan Sprague Dawley, Indianapolis, IN
c DM Purina Veterinary diets, Nestlé-Purina, Vevey, Switzerland
d Careflow, Becton Dickinson, Basel, Switzerland
e L-arginin-hydrochlorid 21%, B-Braun, Sempach, Switzerland
f Trasylol, 500 KIU/ml, Bayer Pharmaceuticals Corporation, Geneva, Switzerland
g Caninsulin, Intervet International BV, Boxmeer, the Netherlands
h Lantus, Sanofi Aventis, Meyrin, Switzerland
i Cobas Integra, Roche, Basel, Switzerland
j Glucose and fructosamine, Roche
k Linco Porcine Insulin RIA Kit, Millipore, Zug, Switzerland
l Glucagon ICN Biomedicals, MP Biomedicals Europe, Basel, Switzerland
mGraphPad Prism 4, GraphPad Software Inc, San Diego, CA
This study was partially supported by a grant from Nestlé Purina PetCare.
- 2Prevalence of feline diabetes mellitus in a feline private practice. J Vet Intern Med 2003;17:433–434., , ,
- 15Diminished insulin secretory response to glucose but normal insulin and glucagon secretory responses to arginine in a family with maternally inherited diabetes and deafness caused by mitochondrial tRNA(LEU(UUR)) gene mutation. Diabetes Care 2001;24:1253–1258., , , et al.