Five to 10 per cent of cats have diabetes classified as “other specific types of diabetes,”^2,⁵ which results from diseases that cause insulin resistance such as acromegaly or hyperadrenocorticism, or from conditions that destroy pancreatic beta cells such as neoplasia or pancreatitis.^6,⁷ The other specific types of diabetes constitute a much greater percentage of the diabetic population in referral practice, especially tertiary referral practice, where most referred diabetic cats are difficult to control and are atypical of the majority of diabetic cats seen in private practice.^2,⁵ Other specific types of diabetes also appear to be more prevalent in developing countries where obesity is rare. Achieving remission is only possible if there are functional beta cells remaining. Therefore cats with other specific types of diabetes mellitus that destroy beta cells, or cats with long-term poorly controlled diabetes mellitus that leads to chronic glucotoxic damage to beta cells, are unlikely to achieve remission, and the goals of therapy will be to control clinical signs.

TYPE 2 DIABETES MELLITUS

In human beings type 2 diabetes mellitus previously was called adult-onset diabetes; however, that name has been discarded because many new cases are in people less than 20 years of age.⁸ It also was called non–insulin-dependent diabetes, but that name also is not recommended because many patients require insulin to achieve good glycemic control. Insulin resistance is a hallmark of type 2 diabetes. Diabetic cats on average are six times less sensitive to insulin than healthy cats, which is a similar magnitude of insulin resistance to human beings with type 2 diabetes.⁹ In human beings insulin resistance predominantly is the result of the sum of the underlying insulin resistance (sensitivity), which is genetically determined, coupled with acquired insulin resistance, which is largely the result of obesity and physical inactivity. A range of insulin sensitivities are present in healthy, ideal-weight cats, and insulin sensitivity also is likely genetically determined in cats.¹⁰ Cats with insulin sensitivities below the population median have three times the risk of developing impaired glucose tolerance with weight gain.¹⁰ Weight gain of 44 per cent in cats over a period of 10 months decreased insulin sensitivity by half.¹⁰ In human beings and dogs physical inactivity leads to insulin resistance independent of body weight, but whether this occurs in cats has not been investigated. Drugs, especially long-acting or repeated glucocorticoid administration, induce insulin resistance and are a frequent precipitator of clinical signs of diabetes in cats. Hyperglycemia also induces insulin resistance and is reversible with improved glycemic control.

Loss of beta cell function is another hallmark of type 2 diabetes. A major cause of beta cell loss in type 2 diabetes is thought to be apoptosis triggered by beta cell damage.¹¹ Beta cell damage is thought to be associated with chronic hyperfunction that occurs secondary to chronic insulin resistance. Other causes of loss of beta cells include pancreatitis and islet amyloid deposition.^2,⁵ Some loss of function is reversible, as occurs in the early stages of glucose toxicity. Loss of beta cells occurs in the later stages of glucose toxicity, resulting in irreversible loss of beta cell secretion. Logically the greater the loss of beta cells, the lower the probability for remission will be in an individual cat.

DIAGNOSIS OF DIABETES MELLITUS

In human beings type 2 diabetes is diagnosed when blood glucose is persistently above 126 mg/dL (7 mmol/L).¹² In cats diabetes usually is only diagnosed when blood glucose concentration is above the renal threshold, causing obligatory water loss and hence the signs of polyuria and polydipsia. This generally is associated with a blood glucose of 234 to 288 mg/dL (13 to 16 mmol/L) or above.¹³ In cats no epidemiological studies have been performed to demonstrate the adverse health effects of persistent mild to moderate hyperglycemia, for example, between 144 and 252 mg/dL (8 and 14 mmol/L). In human patients the cutoff blood glucose concentration for diabetes mellitus has been lowered consistently over time as more information has become available on the adverse effects of mild hyperglycemia, including microvascular damage and retinopathy. Similar studies need to be performed in cats. It is likely that if cats were classified as diabetic with persistent fasting blood glucose concentrations of 180 mg/dL (10 mmol/L) or greater, most diabetic cats also would be non–insulin dependent and a much greater proportion could be controlled with weight loss and diet alone.

Measuring serum fructosamine is not a sensitive indicator of persistent hyperglycemia in cats. In healthy cats infused with glucose to maintain blood glucose concentrations at 306 mg/dL (17 mmol/L) for 6 weeks, serum fructosamine concentration was not consistently above the upper limit of the reference range.¹⁴ Serum fructosamine assay is most useful in situations in which there is very poor clinical history and it is unknown whether hyperglycemia is associated with stress, or if it is truly persistent and associated with diabetes mellitus. However, this could be answered more simply, quickly, and less expensively by hospitalizing the cat and measuring blood glucose at least 3 hours later. Provided struggling does not occur with the second blood collection, blood glucose concentration should be in the normal range when measured 3 to 6 hours later, if hyperglycemia is caused by stress associated with the first blood collection. If blood glucose concentration is still mildly increased (180 to 252 mg/dL; 10 to 14 mmol/L), then measure it again the following morning to clarify the diagnosis. If blood glucose concentration is greater than 270 mg/dL (>15 mmol/L), institute therapy for diabetes immediately. If in doubt about the diagnosis, err on the side of diagnosing diabetes and institute appropriate treatment and monitoring of blood glucose concentration.

GLUCOSE TOXICITY

Glucose toxicity is an important phenomenon to understand because it relates to treatment goals, remission, and choice of insulin. Glucose toxicity describes the suppression of insulin secretion from beta cells secondary to prolonged hyperglycemia.^15,¹⁶ We have shown that insulin secretion in healthy cats is suppressed to concentrations found in insulin-dependent diabetic cats within three to seven days of blood glucose concentrations being maintained in the range of 540 mg/dL (30 mmol/L).¹⁶ Glucose toxicity is dose-dependent, and less suppression occurs at lower glucose concentrations. Glucose toxicity is particularly important when superimposed on hyperfunctioning beta cells that are already compromised because of loss of beta cell mass from amyloid deposition or pancreatitis.¹⁷ In this situation of hyperfunctioning beta cells, even low levels of hyperglycemia cause further rapid deterioration of beta cell function, worsening hyperglycemia, and eventually signs of overt diabetes mellitus.¹⁷ Initially glucose toxicity causes reversible suppression of insulin secretion from beta cells, but later it causes irreversible loss of beta cells.

Achieving resolution of glucose toxicity is critically important to the goal of achieving diabetic remission. This is highlighted by data demonstrating that cats changed to a protocol to achieve excellent glycemic control within 6 months of diagnosis are much more likely to achieve non–insulin-dependent status than cats whose glucose concentrations are not strictly controlled until longer than 6 months after diagnosis.⁴

PREDICTORS OF DIABETIC REMISSION

In a study of 55 diabetic cats treated with glargine, only four variables were found to be associated with a higher probability of remission.⁴ They were early implementation of intensive blood glucose control, prior glucocorticoid treatment, absence of signs of neuropathy, and lower maximal insulin dose. For cats who started intensive blood glucose control within 6 months of diagnosis of diabetes, the remission rate was 84 per cent. The remission rate decreased to 35 per cent for cats who were started on the same protocol more than 6 months after diagnosis; the difference was highly significant (p = 0.0002). Therefore it is critically important that excellent glycemic control be achieved sooner rather than later.

Cats treated with glucocorticoids in the 6 months prior to being diagnosed with diabetes were more likely (p = 0.001) to go into remission than cats without prior treatment with these drugs. The addition of marked insulin resistance secondary to glucocorticoid administration likely precipitates an acute onset of signs, which probably results in treatment being sought earlier, and hence earlier resolution of glucose toxicity than if the onset of signs had been slow and insidious.

Cats displaying a plantigrade stance at diagnosis or milder signs of peripheral neuropathy, such as a difficulty climbing stairs, were significantly less likely (p = 0.02) to go into remission. It is probable that cats with signs of neuropathy had uncontrolled hyperglycemia longer than cats without neuropathy, and hence greater loss of beta cells from the effects of prolonged glucose toxicity, and lower remission rates.

The mean maximum insulin dose required was lower in cats who became non–insulin-dependent during the study (0.43 IU/kg q12h) compared with cats who remained insulin-dependent (0.66 IU/kg q12h). Other factors examined that were not found to be predictors of remission were age at diagnosis, gender, weight at diagnosis, evidence of diabetic ketoacidosis at diagnosis, presence of chronic kidney disease or hyperthyroidism, and frequency of asymptomatic hypoglycemia. Obesity was not correlated negatively with remission.

INSULIN CHOICE

Because remission should be the initial major goal of therapy in newly diagnosed cats, the protocol selected must provide the cat the greatest probability of achieving this outcome. The intermediate-acting insulins such as neutral protamine Hagedorn (NPH) and Lente insulin have too short a duration of action to control blood glucose satisfactorily when given q12h. For porcine Lente insulin (Vetsulin, Caninsulin) given q12h, we have shown that typically for approximately 4 hours before each insulin dose there is minimal glucose lowering effect from the insulin, which translates to minimal insulin action for 8 hours out of every 24.¹⁸ Although this duration of action is sufficient to control clinical signs in many cats, the marked hyperglycemia (>324 mg/dL; >18 mmol/L) twice daily continues to suppress and damage beta cells, leading to significantly lower remission rates compared with that achieved with longer-acting insulin.³ NPH insulin tends to have a shorter duration of action than porcine Lente insulin, and is not recommended for use in cats. Although protamine zinc insulin (PZI) has a longer duration of action than Lente insulin, it appears to be associated with lower remission rates than the new long-acting insulin analogues glargine (Lantus, Aventis) and detemir (Levemir, Novo Nordisk).^3,¹⁹ In a small study of newly diagnosed diabetic cats, two of eight cats achieved remission with porcine Lente insulin, three of eight with PZI, and eight of eight with glargine.³ Further evidence that glargine produces superior remission rates to Lente insulin is found in a recent study of 50 diabetic cats treated previously for a median of 16 weeks with other insulins, predominantly porcine Lente insulin, who failed to achieve remission. After changing to glargine and a protocol of intensive glucose monitoring, remission rates of 84 per cent were achieved if cats were started on the intensive protocol within 6 months of diagnosis.⁴ In some countries veterinarians are required legally to use insulin registered for veterinary use as a first line of therapy. Because remission rates for cats are significantly lower if intensive glucose control is not instituted early, if diabetic remission does not occur within 4 to 8 weeks of using an insulin registered for veterinary use, it is recommended that the insulin be changed to glargine or detemir to facilitate remission.

Based on reported data the first choice of insulin in newly diagnosed diabetic cats is glargine. Early studies of limited numbers of cats indicate that there is little clinical difference between glargine and detemir.¹⁹ Studies in healthy cats suggest that detemir has a longer duration of action than glargine and is perhaps the insulin of choice in cats.²⁰ Because there is substantially greater experience with the use of glargine in cats, and the pharmacokinetics and dynamics of glargine, but not detemir, have been reported in cats, glargine currently is the insulin of choice until more data are available. In the study of healthy cats detemir had a longer duration of action (median 800 minutes; range 525 to 915) than glargine (median 470 minutes; range 295 to 950 minutes) and less variation in duration of action between cats.²⁰ In human patients detemir also has been shown to be more consistent in its duration of action than glargine. Therefore, in cats in whom glargine appears to have too short a duration of action with q12h administration, detemir should be trialed. In countries where there is a legal obligation to first use a veterinary-use insulin, PZI would be the first-choice veterinary insulin if available, and porcine Lente insulin would be the second-choice veterinary insulin. It is recommended that if remission is not achieved within 4 to 8 weeks, the insulin be changed to glargine or detemir. It is likely that if excellent glycemic control can be achieved early and episodes of marked hyperglycemia avoided, the loss of beta cells associated with glucose toxicity is minimized and remission is facilitated.