Diabetes Mellitus

Diabetes Mellitus

Symposium on Endocrinology Diabetes Mellitus William D. Schall, D.V.M.,* and Larry M. Cornelius, D.V.M., Ph.D.t Diabetes mellitus is one of the most...

3MB Sizes 0 Downloads 17 Views

Symposium on Endocrinology

Diabetes Mellitus William D. Schall, D.V.M.,* and Larry M. Cornelius, D.V.M., Ph.D.t

Diabetes mellitus is one of the most commonly encountered metabolic disorders of dogs and cats. Its rate of occurrence in these two species is not precisely known, but it is generally agreed to be more common in the dog. Diabetes has been documented in dogs from less than one year to over 15 years of age and in cats from 2 to 15 years of age. More then 75 per cent of dogs and cats are older than 5 years of age when the diagnosis is established. Dachshunds, poodles, and Siamese breeds may be at increased risk. 20 · 30 Bitches are at increased risk over dogs, but queens are at equal risk when compared with toms. 20 · 30 The inheritability of diabetes mellitus in the dog and cat has not been established, but a few related dogs have been documented to have the disorder.

PATHOGENESIS For more than 50 years diabetes mellitus has been considered to be a monohormonal disease caused by a relative or absolute lack of insulin. New information, however, supports the hypothesis that diabetes is a bihormonal disease in which insulin lack is accompanied by glucagon excess. The excessive glucagon plays a major role in altering intermediary metabolism that leads to hyperglycemia and ketoacidosis.33 Glucagon is manufactured by alpha-2 cells within the pancreas and normally acts to increase blood glucose concentration if the concentration becomes too low. This hyperglycemic effect of glucagon is due to its influence on hepatic glycogenolysis and gluconeogenesis. *Associate Professor, Department of Small Animal Surgery and Medicine, Michigan State University College of Veterinary Medicine, East Lansing, Michigan t Associate Professor, Department of Small Animal Medicine and Surgery, University of Georgia College of Veterinary Medicine, Athens, Georgia

Veterinary Clinics of North America- Vol. 7, No. 3, August 1977

613

614

WILLIAM

D.

ScHALL, LARRY

M.

CoRNELIUS

Insulin is manufactured by beta cells of the pancreas and normally acts to lower blood glucose concentration. Gluco!>e enters cells more readily, glycogen synthesis is increased, and glycogenolysis and gluconeogenesis are inhibited in response to physiological amounts of insulin. Normally, these two hormones maintain blood glucose concentrations within narrow limits through feedback mechanisms. The most important stimulus for insulin synthesis and release is increased blood glucose concentration and inhibition of insulin release normally occurs at blood glucose concentrations below 50 mg per dl. Glucagon release in contrast is normally stimulated in hypoglycemic states and its secretion is decreased when hyperglycemia occurs. Other compounds and hormones can effect the release of insulin and glucagon, but none is as important as glucose. Most states of endogenous hyperglycemia in man and animals are now known to be accompanied by hyperglucagonemia regardless of plasma insulin concentration. Unger and Orci33 have reviewed investigations, the results of which equate endogenous hyperglycemia with hyperglucagonemia. Spontaneous diabetes in man and Chinese hamsters, alloxan diabetes in dogs and rats, and stress hyperglycemia and diabetes secondary to pancreatectomy in dogs are included. In pancreatectomized dogs, the source of glucagon is the alpha cells of the gastric fundus and duodenum. The extra pancreatic glucagon is biologically, immunologically, and physiochemically identical to pancreatic glucagon. The administration of pharmacological doses of somatostatin (growth hormone release-inhibiting factor) suppresses both insulin and glucagon secretion. Through the use of this tool, it is now known that endogenous hyperglycemia does not occur if secretion of both hormones is suppressed, but does occur if exogenous glucagon is infused in the bihormonally suppressed animal. It has also been established that somatostatin administration prevents the occurrence of ketoacidosis in human diabetic 'patients deprived of insulinP Although plasma glucagon concentration has not been determined in cats or dogs with naturally occurring diabetes mellitus, available data suggest that it is unlikely that either hyperglycemia or ketoacidosis occurs in any species of animal in the absence of hyperglucagonemia. The biochemical abnormalities that characterize diabetes mellitus and diabetic ketoacidosis are now understood on the basis of biharmona! control of substrate metabolism. Hyperglycemia of diabetes results primarily from excess glucose production rather than underutilization. Once glucogen stores have been depleted, excess glucose production is due to increased gluconeogenesis. The rate of gluconeogenesis is determined by the rate of amino acid release from muscle and uptake by the liver. Although the release of gluconeogenic amino acids from muscle occurs because of insulin deficiency, glucagon en-

615

DIABETES MELLITUS

hances hepatic uptake and appears to be more important in determining the rate of gluconeogenesis. 33 Glucagon plays a similar role in the development of ketonemia characteristic of diabetic ketoacidosis. The release of free fatty acids in diabetes is primarily a result of insulin deficit, but hepatic uptake and conversion to ketone bodies occurs because of glucagon excess. Strong evidence supports the hypothesis that hyperglucagonemia may be more important than hypoinsulinemia in the pathogenesis of diabetes mellitus. Evidence to the contrary does exist, however, and has been summarized by Levine. 19 Although still a subject of debate, the concept that excess glucagon plays an important role in diabetes has obvious therapeutic implications. Current therapy consists of insulin replacement, but suppression of glucagon production or release may become important in the future. 13

LABORATORY EVALUATION Establishing the Diagnosis The signs of diabetes mellitus and its complications have been well described. 2 • 4 Tests to detect the presence of excessive glucose in the blood and urine are the most useful in establishing a diagnosis of diabetes mellitus. Ketonemia and ketonuria may accompany hyperglycemia and glucosuria and are further evidence of the diabetic state. Several methods are available to measure blood and urine glucose. Rapid detection methods utilizing reagent-incorporated sticks and tablets are generally reliable in establishing a diagnosis. Although published studies have indicated some inaccuracy in these methods, 16 • 31 blood and urine glucose levels are generally high by the time the animal is brought to the hospital. Use of these tests will usually allow a presumptive diagnosis and the rapid initiation of therapy (if necessary) while the more accurate procedures are being performed in the laboratory. Some precautions must be observed in the use of the semiquantitative methods. Dextrostix* contains the enzyme glucose oxidase and is used for the semiquantitative estimation of whole blood glucose. As with any determination based upon an enzyme-substrate interaction, care must be taken to ensure proper ambient conditions. The use of outdated strips (greater than 6 months old) should be avoided. The strips should be stored and used at room temperature. In the actual determinations, it is desirable to eliminate as many variables as possible. The same person should perform all tests if practical. Meticulous ad*Ames Company, Elkhart, Indiana.

616

WILLIAM

D.

SCHALL, LARRY

M.

CORNELIUS

herence to the manufacturer's instructions should be routine. Also, in the comparison of colors on the stick with the accompanying color chart, similar lighting conditions should always be present. The use of an improved reflectance meter system* with a synthetic whole blood controlt to read Dextrostix results has been reported to improve the accuracy of the method. 6 • 31 Urine glucose is determined semiquantitatively by measuring reducing substances (Clinitest tablets:j:) or by glucose oxidase strips (Ketodiastix,:j: Clinistix,:j: Labstix,:j: Tes-Tape§). Clinitest tablets are not specific for glucose, and other reducing substances such as vitamin C1 and salicylates7 may cause false positive reactions. False negative results may occur as a consequence of vitamin C inhibition of the glucose oxidase methods. 1 In one reported study, Clinitest tablets (2 drop and 5 drop method), Tes-Tape, and Ketodiastix were evaluated using urine-containing glucose in low(~ 376 mg per 100 ml), medium (376 to 1500 ml per 100 ml), and high (~ 1500 mg per 100 ml) quantities. At high urine glucose concentrations, the 2-drop Clinitest method gave better quantitation than the other methods. At intermediate levels, there was little difference between methods, whereas at low urine glucose concentrations, the 2-drop Clinitest method was often insensitive. In the low range, the other methods gave comparably accurate results, with the exception that Tes-Tape was sometimes positive with normal levels of glycosuria ( 1 to 15 mg per 100 ml). 16 The nitroprusside reaction is the basis for semiquantitative estimation of ketones in serum or plasma (Acetest tablets*) and urine (Acetest tablets, Ketodiastix). This reaction measures acetoacetic acid and acetone but not beta-hydroxybutyrate. ~ In human diabetic ketoacidosis accompanied by lactic acidosis, a weakly positive reaction may not accurately reflect the magnitude of ketosis. This is because the altered redox state favors the conversion of acetoacetate to beta-hydroxybutyrate. This may result in an increase in the ratio of beta-hydroxybutyrate to acetoacetate from the normal of 3:1 up to as high as 15: l, thus causing underestimation of the severity of ketosis. 8 • 23 Although this phenomenon has not yet been documented in diabetic animals, the possibility certainly exists. Acetest was found to be a more sensitive and accurate indicator or urinary ketones than Ketodiastix. 16 For plasma or serum ketones, the Acetest tablet should probably be crushed before use to allow better distribution and mixing of the reagent and sample. In severe ketosis, serial dilutions of plasma or serum, such as 1:2, 1A, 1:8, 1:16, and 1:32, can be made with saline or control serum. One then records the highest dilution at which a positive reaction is observed. *Eyetone/Dextrostix- Ames Company, Elkhart, Indiana. tTek-Chek Synthetic Control-Ames Company, Elkhart, Indiana. *Ames Company, Elkhart, Indiana. §Eli Lilly Company, Indianapolis, Indiana.

617

DIABETES MELLITUS

Because of convenience, it is common practice in many laboratories to use one of the dip-stick methods to screen urine samples for glucose and ketones. Considering the above factors, however, it would appear that Clinitest and Acetest tablets are the preferred methods whenever more accurate semiquantitation is desired. This latter situation is common when trying to finely regulate a diabetic animal on insulin in the home environment. When measuring and interpreting blood glucose values, a few precautions are necessary. Blood sampled for glucose determination should not be allowed to stand for more than 30 minutes since erythrocyte glycolysis will utilize glucose and cause significant reductions in the value obtained. If immediate analysis is impractical, the sample should be centrifuged and the serum removed and frozen for later analysis. Alternatively, sodium fluoride may be used as the anticoagulant with the fluoride also acting to reduce blood enzyme activity. Only fasting samples (at least 12 hours) should be taken. Hyperglycemia is common after meals. Also, hyperlipemia can cause dramatic increases in blood glucose values obtained by some laboratory methods. Excessive fright, excitement, and/or struggling during sampling can cause hyperglycemia due to epinephrine release and subsequent glycogenolysis. It has been our experience that the stress of debilitating illness in cats is commonly associated with transient hyperglycemia and glucosuria. Normal fasting blood glucose values seldom exceed 120 mg per 100 ml. Most diabetic animals have a blood glucose concentration greater than 100 mg per 100 ml. Animals with fasting values persistently between these two concentrations probably should be given a glucose tolerance test. A procedure utilizing a high dose intravenous glucose tolerance method has been reported to be reliable in dogs. 14 • 34 In this procedure a zerotime fasting blood sample is taken, and 50 per cent glucose (1 gm glucose per kg of body weight) is administered intravenously over a 30 second period. Postinfusion samples of blood are obtained at 5, 15, 30, 45, and 60 minutes for glucose determination. The glucose disappearance coefficient (k) is calculated from the formula: k-

T 26 ~ 3T 1

% decrease of glucose per minute

where T2 - T1 represents the time required for blood glucose to decrease from one arbitrary value to half that value (usually 300 to 150 mg per 100 ml). Values of k which are 2.0 or less are considered evidence of latent diabetes mellitus in dogs. 34 More recently it was reported that a simple k value, calculated from two blood samples taken 5 minutes and 60 minutes after glucose infusion, closely approximated

618

WILLIAM

Table 1.

D.

ScHALL, LARRY

M.

CoRNELIUS

Clinical Differences Between Uncomplicated and Complicated Diabetes Mellitus*

UNCOMPLICATED

COMPLICATED

HISTORY

Polydipsia Polyuria Polyphagia

Anorexia Vomiting Diarrhea Lethargy, weakness Kussmal breathing PHYSICAL EXAMINATION

Alert Normal temperature Cataracts Hepatomegaly

Depression or coma Weakness Dehydration Fever Acetone breath odor Labored breathing Abdominal pain Icterus Shock

*Modified from Cotton, R. B., and Theran, P.: Diabetes mellitus. In Kirk, R. W. (ed.): Current Veterinary Therapy IV. Philadelphia, W. B. Saunders Co., 1971.

the standard k value derived from six blood samples and could be used in clinical situations. 22 The results of intravenous glucose tolerance tests ( 1.1 gm glucose per kg of body weight) in normal and diabetic cats have also been reported. 10 Laboratory Profile in Diabetes Mellitus A number of serious abnormalities may coexist in the diabetic animal. Bacterial infection, acute pancreatitis, chronic renal insufficiency, fatty infiltration of the liver, and pancreatic exocrine insufficiency are a few of the more common disorders that may be present. 3 Other complications which may be present are ketoacidosis, potassium depletion, and hyperosmolal coma. These may or may not be readily apparent from the history and physical examination. Differences in the history and physical examination which may help differentiate complicated and uncomplicated diabetics are listed in Table l. It is extremely important, however, to perform certain laboratory tests to rule out concurrent disorders which, if present, may dramatically affect the mode of therapy and alter the prognosis. Nonketotic Diabetes Without CNS Depression (Uncomplicated). Animals in this category generally are uncomplicated diabetics. If abnormalities are neither reported in the history nor detected upon physical examination, it is always tempting to by-pass laboratory testing (except blood glucose) and proceed with insulin therapy. This approach

DIABETES MELLITUS

619

can have serious consequences in our experience because of subclinical complications which soon become manifest. A suggested laboratory profile for animals in the apparently uncomplicated category includes the following determinations: complete blood count, urinalysis, blood glucose, BUN, SGPT, SAP, total protein (albumin and globulin), sodium. Ketoacidotic Diabetes (Complicated). Ketoacidosis may or may not be associated with abnormal clinical signs. Generally, however, most diabetic animals which are vomiting, lethargic, and dehydrated will be ketoacidotic. These signs should always alert the clinician to the likelihood of serious intercurrent disease and justify a more guarded prognosis. Such patients should be regarded as true medical emergencies and appropriate therapy instituted without delay. A suggested laboratory profile for the ketoacidotic animal includes: complete blood count, urinalysis, blood glucose, BUN, SAP, total protein (albumin and globulin), pH, PC02 , P02 , HC03 , sodium, potassium, and amylase and/or lipase. Without proper laboratory guidance, it is most difficult to achieve a satisfactory outcome in these patients. 21 Diabetic Coma. Animals in coma associated with diabetes mellitus are a medical challenge and require extensive laboratory monitoring. Until recently it has been assumed that severe ketoacidosis was responsible for CNS depression and coma. The results of studies in human diabetics have shown that there may be little correlation between the level of acidosis and coma, but good correlation between serum osmolality and the state of consciousness.U Nonketotic hyperosmolal coma has also been reported in a dog, although the coexistence of acute pancreatitis and azotemia in this patient may have accounted for part or all of the observed CNS depression. 29 Regardless, the laboratory profile of a severely depressed or comatose diabetic should include a serum osmolality determination as well as complete blood count, urinalysis, blood glucose, BUN, SGPT, SAP, total protein (albumin and globulin), sodium, potassium, amylase and/or lipase, pH, PC02 , P02 , and HC03 • Since equipment for measuring serum osmolality is not commonly available, one can instead estimate serum osmolality from the following formula: 18 Serum osmolality= 2 (Na + K) +Blood glucose+ BUN (mOsm per kg) 18 2.8

Results of calculations obtained from using this formula tend to underestimate actual osmolality. Normal serum osmolality is about 300 mOsm per kg, and signs of hyperosmolal coma may occur with an osmolality of 340 or greater. It is our experience, however, that dogs tolerate higher serum osmolalities than people, and signs referable to serum hyperosmolality are not usually seen until the plasma osmolality approaches 375 mOsm per kg. We have had little experience with

620

WILLIAM

D.

ScHALL, LARRY

M.

CoRNELius

feline serum osmolality. It should be remembered that severe CNS depression in a diabetic patient may also be associated with other coexisting disorders.

Monitoring the Response to Therapy- Initial Stabilization Once insulin therapy has been instituted, periodic blood glucose determinations are the most useful and reliable indicators of the efficacy of treatment. Plasma and/or urine ketones can also be measured but probably should not be relied upon because in human diabetics it has been reported that blood acetone and acetoacetate may rise initially in response to insulin therapy due to conversion of beta-hydroxybutyrate to acetoacetate and acetone. 8 Also, acetone is volatile and stored in fat and may be released into the blood and eliminated in the breath and urine long after other blood ketones have returned to normal.2 3 The timing, frequency, and method of the blood glucose determinations are determined by the severity of each case and the type of insulin being administered. It is important to measure the blood glucose near the peak time of insulin action.

THERAPY Therapy of Uncomplicated Diabetes Mellitus Uncomplicated cases generally can be regulated with a long acting insulin such as NPH, which in the dog usually peaks between 8 and 12 hours following subcutaneous administration. In our experience, the NPH insulin peak and duration of action are highly variable in some cats and small dogs, with the maximum effects sometimes being observed as early as 4 hours after subcutaneous administration and the duration of action being only 8 hours. Although the cause of this variability is uncertain, it may be due to the very small quantities of insulin being used in these animals. The blood glucose determinations should be done in the laboratory because Dextrostix is not suitably accurate to allow one to make small adjustments in insulin dosage. It is neither necessary nor desirable to lower the blood glucose concentration into the low to mid normal range while the animal is hospitalized. Rather it is usually easier to achieve final regulation after the patient goes home since diet and exercise may be different. Most clients readily learn to collect urine samples and test for urine glucose and ketones with either Clinitest or Acetest tablets or Ketodiastix. A consistently negative glucose value at the time that NPH insulin peaks and a slight trace reaction 24 hours after administration is probably ideal.

DIABETES MELLITUS

621

Therapy of Diabetic Ketoacidosis Therapy of diabetic ketoacidosis involves the simultaneous titration of fluids, electrolytes, regular insulin and, in some instances, alkalizing agents. Interaction of the biochemical effects of fluids, electrolytes and insulin, in turn, modifies the need for each of these agents. As a consequence of this intricate interaction o(agents contemporaneously administered, repeated reassessment of the biochemical status of the patient is necessary. Immediate Consideration. It is important that patency of the airway be established in those patients presented in a coma. The combination of emesis and sluggish or absent reflexes may necessitate endotracheal intubation. Aseptic catheterization of a vein of one of the extremities provides an immediate means of fluid administration. Jugular catheterization is necessary, however, if central venous pressure is to be monitored. An indwelling catheter should be placed in the urinary bladder so that an oliguric or anuric state can be readily detected. Water. Dehydration can best be assessed by combining the results of physical examination, packed cell volume, and the plasma protein. The extent of water loss usually approaches an amount equivalent to 10 to 15 per cent of body weight. Although rehydration is usually scheduled over 48 hours, rapid restoration of circulating volume is of paramount importance; 50 per cent of the calculated deficit should be administered intravenously in the first 12 hours and fluids should be adminstered at the rate of 90 ml per kg of body weight per hour for the first one or two hours or until the circulating volume is restored. The fluid of choice is a subject of debate. We prefer lactated Ringer's solution but some clinicians use 0. 9 per cent sodium chloride and add supplemental potassium chloride; others use hypotonic fluids. Electrolytes. The diabetic ketoacidotic patient usually is defiicient in sodium potassium but serum sodium and potassium determinations often do not reflect the magnitude of the loss because the deficit may be masked by free water loss. In addition, the extent of potassium depletion is not reflected in serum concentration because addition of this cation to extracellular water occurs at the expense of intracellular concentration. In spite of these limitations it is essential that the serum concentration of sodium and potassium be determined before and during fluid and electrolyte therapy. Hypokalemia is occasionally detected prior to therapy but more often occurs as a result of insulin administration. Decrease in serum potassium concentration as a result of insulin therapy usually is detected concomitantly but may precede decrease in blood glucose concentration. It has been suggested that insulin may facilitate intracellular relocation of potassium independently of its action of glucose transport. Because the degree of hypokalemia may be sufficient to cause

622

WILLIAM

D.

ScHALL, LARRY

M.

CoRNELIUS

death and because uncertainty exists with respect to the time of development of hypokalemia, periodic assessment of serum potassium is important during the first 8 hours of insulin therapy. If serum potassium concentration of less than 3 mEq. per liter is detected, especially if the patient is not fully hydrated, cautious replacement therapy is indicated. Potassium is administered intravenously to the ketoacidotic cat or dog because oral intake is usually precluded. Potassium chloride is usually added to lactated Ringer's solution at the rate of 30 mEq per liter, making a solution with 34 mEq per liter. The rate of intravenous potassium administration should not exceed 0.5 mEq per liter of body weight per hour. In the case of a solution containing 34 mEq per liter, only 15 ml per kg of body weight per hour should be administered. Lactate and Bicarbonate. It has been our experience that neither clinical signs nor laboratory data, other than actual blood gas determination or the equivalent, adequately predict the acid-base status of diabetic ketacidotic patients. Although the clinician may elect not to treat mild or moderate metabolic acidosis, we think it essential that severe metabolic acidosis be identified and treated. Documentation of severe acidosis can be accomplished only through laboratory determinations. It is ideal to measure PC02 and pH. Bicarbonate can then be estimated from standard nomograms. If this equipment is not available, other methods may be used. The C02 content of the plasma can be measured rapidly, accurately, and relatively inexpensively by using a commercially available total C02 apparatus*. 5 The C02 content is largely comprised of bicarbonate and thus is a measure of the metabolic component of acid-base balance. If neither pC02 and pH nor total C02 can be determined, the dissipation of plasma and urine ketones generally can be interpreted as evidence of improvement of the metabolic acidosis, as can a rising urine pH. The total amount of lactate or bicarbonate needed to correct the acidosis can be estimated by multipying the bicarbonate deficit (normal serum concentration minus measured) times one-half the body weight in kg. The calculated quantity of base should be administered over-48 hours to avoid paradoxical CSF acidosis which is discussed under the section on diabetic comas. Consideration should be given to the alkalinizing agent chosen. Acetate may be ketogenic and for this reason is seldom used for diabetic acidosis. If lactic acidosis complicates diabetic ketoacidosis then the usual alkalinizing effect of lactate will not occur, although there is no evidence that sodium lactate will contribute to lactic acidosis. No argument exists regarding alkalinizing potential of bicarbonate, but correction of metabolic acidosis should be undertaken over-48 hours. Insulin. Regular insulin is used for treating diabetic ketoacidosis. All or part of the insulin is administered intravenously because absorp*Harleco C02 Apparatus- Harleco Division, American Hospital Supply, Philadelphia, Pennsylvania.

DIABETES MELLITUS

623

tion from subcutaneous sites is poor owing to poor tissue perfusion secondary to dehydration. In the past, regular insulin has been administered as a bolus and repeated every 2 hours until the plasma glucose concentration, which was determined at 2 hour intervals, was less than 200 mg per dl. NPH insulin was then administered subcutaneously. The initial doses of insulin used has varied widely. Some clinicians have advocated the use of high initial insulin doses based on the notion that ketosis or low blood pH impairs tissue insulin responsiveness. There is no evidence, however, to support this contention. Low doses of insulin have been documented to be just as effective as high doses in correcting ketoacidosis in man. 32 When regular insulin is injected as an intravenous bolus the dose for the dog is 1 to 2 units per kg (0.5 to 1.0 unit per lb) of body weight. The lower dose on a per weight basis is used in larger dogs and the higher dose in used for smaller dogs. Cats are insulin-sensitive compared with dogs. For this reason the recommended dose is 0.5 units per kg of body weight. Intermittent bolus intravenous insulin administration as described has been the standard therapeutic approach to diabetic ketoacidosis in man and dogs for years. Several papers, however, report the efficacy of continuous infusion of insulin at a low dose for treatment of human diabetic ketoacidosisY We have used slow intravenous insulin infusion to treat six diabetic ketoacidotic dogs and find it an appealing alternative to intermittent intravenous insulin. The advantages of low dose infusion in man compared with intermittent intravenous injection have been reported. 24 Both hypokalemia and hypoglycemia are less likely to occur and, if either of these complications should arise, stopping the insulin infusion will result in a rapid decline of the serum insulin concentration. Second, the timing and dose of subsequent insulin injections are not problems. The insulin infusion is continued until the blood glucose has reached a satisfactory concentration at which time standard administration of NPH insulin is begun. Finally, the method is simple and relatively inexpensive in that the only spe\ial item of equipment needed is a pediatric drip set to more finely regulate the slow drip rate. We add 5 units of regular insulin to 500 ml of lactated Ringer's solution and, using a pediatric infusion set, administer insulin at the rate of 0.5 to 1.0 units per hour. To achieve adequate rehydration, noninsulin-containing fluid is administered simultaneously at an appropriate rate. The plasma glucose concentration is determined every one to two hours and the insulin infusion stopped when the blood glucose is 200 mg per dl or less. NPH insulin is then administered. All of the six diabetic ketoacidotic dogs that we have treated in this manner have had blood glucose concentrations of 200 mg per dl or less within 8 hours after the infusion was started. We are currently investigating the efficacy of the low dose infusion regimen by prospective randomized comparison with conventional intermittent intravenous treatment program.

624

WILLIAM

D.

ScHALL, LARRY

M.

CoRNELIUS

The low total doses of insulin that appear to be effective in reversing diabetic ketoacidosis are probably even less than calculated because up to 75 per cent of the insulin adheres to the fluid bottle and infusion apparatus. 26 The use of somatostatin to inhibit glucagon secretion in combination with insulin administration, to our knowledge, has not been attempted in dogs or cats with naturally occurring diabetic ketoacidosis, but initial studies suggest that it may be a useful adjunct in man. 13 • 28

COMPLICATIONS Pancreatitis

Pancreatitis complicates diabetic ketoacidosis more frequently than any other disease and is the most common cause of death. 20 Increased serum activity of either amylase or lipase greater than two times normal in a vomiting patient is strongly suggestive of pancreatitis. Lesser increases in the serum activity of either enzyme may be the result of dehydration and consequent decreased glomerular filtration rate rather than pancreatic inflammation. Diabetic ketoacidotic patients with pancreatitis should be treated for both diseases simultaneously. For pancreatitis all oral intake should cease and maintenance fluid requirements should be administered by the intravenous and subcutaneous routes. Oral intake is not allowed until serum lipase and/or amylase activity are normal and the patient no longer is vomiting (usually 3 to 4 days). To decrease pancreatic secretory activity, atropine sulfate is administered subcutaneously four times daily at a dose of 0.04 to 0.08 mg per kg (0.02 to 0.04 mg per pound) of body weight. Because the indication for antibiotics, analgesics, tranquilizers, and glucocorticoids is equivocal, we do not routinely administer them. Hypocalcemia occasionally occurs and if documentedshould be treated. Pancreatic Exocrine Insufficiency

·Exocrine insufficiency occasionally complicates diabetes mellitus. It is seldom of immediate concern in ketoacidotic patients but must be treated on a long-term basis for adequate regulation. Documentation is usually simple because associated steatorrhea is severe and many fat droplets are evident microscopically after feces are stained with Sudan IV. Other evidence suggestive of pancreatic exocrine insufficiency is failure of gelatin digestion, multiple starch granules evident when feces are stained for starch, and lack of postprandial lipemia without but not with the addition of exogenous pancreatic enzymes to a high fat meal. Twenty-four hour fecal fat determinations usually reveal that from 25 to 55 per cent of ingested fat can be recovered in feces. 15 Considerable variation exists between enzyme replacement products, some commercial preparations having been documented as supe-

DIABETES MELLITUS

625

rior. 25 Substantial variation in enzyme content between various lots of a given product also has been demonstrated and for this reason it is necessary to adjust the amount of replacement until feces are grossly normal. Even under optimal conditions, some degree of maldigestion persists. We believe powder to be superior to tablets or capsules and recommend a 20 minute preingestion incubation period. Azotemia Dogs and cats with diabetic ketoacidosis are often azotemic as a result of dehydration. If, after adequate fluid replacement, the azotemia persists then consideration must be given to azotemia of renal origin. Fatty Metamorphosis Hepatomegaly is common in diabetic cats and dogs and is usually due to fatty metamorphosis. Moderate increase in serum transaminase activity is often present. Increased serum alkaline phosphatase activity is sometimes documented and when present is often profound. Increased BSP retention that reverted to normal l to 2 months following insulin therapy has been reported in dogs. 20 Diabetic Comas Although coma is not common in diabetic patients, its occurrence is important because of multiple potential causes. Ketoacidosis. Coma can occur in diabetic ketoacidosis because of the metabolic acidosis alone. The occurrence is not common because in metabolic acidosis, cerebral spinal fluid (CSF) bicarbonate concentration is well preserved and CSF acidosis rarely occurs. Respiratory acidosis, in contrast, is often accompanied by stupor or coma because C02 can easily diffuse into CSF resulting in CSF acidosis. Spontaneous lactic acidosis has been documented in human patients with diabetes mellitus and may occur in animals. It should be suspected if severe metabolic acidosis in a diabetic patient is documented without commensurate ketonemia because lactic acidosis is associated with increased beta-hydroxybutyrate and decreased acetoacetic acid and acetone. The nitroprusside reaction, the usual basis for ketone estimation, does not detect beta-hydroxybutyrate. Paradoxical CSF AcidosisY Stupor leading to coma can occur as a result of rapid correction of metabolic acidosis with intravenous alkalinizing agents. The phenomenon occurs because the CSF becomes more acid at a time when blood pH is corrected to normal. In metabolic acidosis CSF pH is better preserved than is blood pH because of CSF bicarbonate retention. If blood pH is rapidly corrected, however, blood bicarbonate and. pH may exceed CSF concentrations. Blood C02 content then increases to a concentration greater than that of CSF and rapidly diffuses into CSF before bicarbonate can do likewise. The consequence is decreased CSF pH.

626

WILLIAM

D.

ScHALL. LARRY

M.

CoRNELIUS

Hyperosmolal Coma. Coma can occur as a result of increased serum osmolality. Patients are usually but not always nonketotic. The hyperosmolality occurs as a result of severe hyperglycemia and hypernatremia and can be determined by use of an osmometer or estimated from the formula. Normal canine osmolality is about 300 mOsm per kg of serum; hyperosmolal coma should be suspected whenever the serum or plasma osmolality of a diabetic dog or cat exceeds 340 mOsm and coma cannot be otherwise explained. Hypotonic rather than isotonic fluids should be used in the therapy of hyperosmolal diabetic coma. Fluids such as half-strength saline, quarter-strength saline in 2.5 per cent dextrose or commercial maintenance solution can be used depending upon availability and the relative contribution of sodium and glucose to hyperosmolality. The amount of dextrose in quarter-strength saline in 2.5 per cent dextrose will contribute little to the glucose burden and this fluid has the desired hypotonic property. Maintenance solutions contains ketogenic acetate but usually can be used if it is the only available hypotonic fluid and the hyperosmolality is not complicated by ketosis. Cerebral Edema. There is ample documentation of instances in which human patients have lapsed into comas and died in association with increased intracranial pressure and cerebral edema while being treated for diabetic ketoacidosis. The mechanism of the cerebral edema is currently a subject of debate and investigation. One hypothesis involves the polyol pathway and a pathogenesis similar to that proposed for diabetic cataract development. Increased intracranial pressure has been observed in dogs after experimental elevation of blood glucose above 400 mg per I 00 ml for 4 hours and subsequent rapid decrease in blood glucose concentration as the animals were rehydrated with isotonic saline. Hypoglycemic Coma. The known diabetic animal receiving insulin administered by the pet owner may be presented in a coma to the clinician. In spite of a carefully taken history, it may not be immediately apparent whether the coma is the result of hypoglycemia or ketoacidosis. In this instance rapid blood glucose determination by use of Dextrostix or other laboratory methods is necessary to establish the cause of coma. Urine glucose is potentially misleading in that the urine tested may have been formed hours earlier and thus does not reflect the current status of the patient. In the absence of resources necessary for rapid blood glucose determination, the intravenous injection of hypertonic dextrose (I 0 to 25 ml of 20 per cent dextrose) can help establish the cause of coma. The hypoglycemic animal should respond almost immediately, but the relatively small amount of dextrose will not have an appreciably adverse effect on the ketoacidotic patient. Cataracts

Cataracts are a common complication of diabetes mellitus in dogs

627

DIABETES MELLITUS

but have not been reported in cats. Cataract removal is often successful but deaths associated with anesthesia and surgery are not uncommon in diabetics. The risks of the procedure must be compared with the potential benefits and discussed with the dog's owner.

Vascular Disease Microvascular complications and atherosclerosis are common in human diabetic patients. The rarity of these complications in cats and dogs is probably a reflection of shorter life span compared with man inasmuch as vascular complications are seldom diagnosed until human diabetics have been so for five years or more. Euthanasia The most common reason for euthanasia has been reported to be unwillingness on the part of pet owners to continue treatment. 20 This suggests that client education and understanding are important obligations for the veterinarian.

REFERENCES I. Brandt, R., Guyer, K. E., and Banks, W. K.,Jr.: A simple method to prevent vitamin

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

C interference with urinary glucose determinations. Clin. Chim. Acta, 51:103, 1974. Capen, C. C., Belshaw, B. E., and Martin, S. L.: Endocrine disorders. In Ettinger, S. ]. (ed.): Textbook of Veterinary Internal Medicine. Philadelphia, W. B. Saunders, 1975. Cotton, R. B., Cornelius, L. M., and Theran, P.: Diabetes mellitus in the dog: A clinicopathologic study. J.A.V.M.A., 159:863, 1971. Cotton, R. B., and Theran, P.: Diabetes mellitus. In Kirk, R. W. (ed.): Current Veterinary Therapy IV. Philadelphia, W. B. Saunders, 1971. Creno, R.]., Farrand,]., and Koch, B. H.: An evaluation of the Harleco total carbon dioxide apparatus. Am.]. Med. Tech., 40:50, 1974. Davis, A. E.: Evaluation of an improved reagent strip system for measuring blood glucose. Am.]. Med. Tech., 42:18, 1976. Duncan,]. R., and Prasse, K. W.: Urinalysis. In Duncan,]. R. and Prasse, K. W.: A syllabus of veterinary clinical pathology. University of Georgia, College of Veterinary Medicine, Athens, Georgia, 1973. Felig, P.: Diabetic ketoacidosis. New Engl.]. Med. 290:1360, 1974. Felig, P.: Insulin: Rates and routes of delivery. New Engl.]. Med., 291: l 031, 1974. Finn,]. P., Martin, C. L., and Manns,]. G.: Feline pancreatic islet cell hyalinosis associated with diabetes mellitus and lowered serum insulin concentrations. J. Small Anim. Pract., 11:607, 1970. Fulop, M., Tannenbaum, H., and Dreyer, N.: Ketotic hyperosmolar coma. Lancet, 2:635, 1973. Gerich,]. E., Lorenzi, M., Bier, D. M., et al.: Prevention of human diabetic ketoacidosis by somatostatin. Evidence for an essential role of glucagon. New Engl. ]. Med., 292:985, 1975. Gerich,]. E., Lorenzi, M., Karam,]. H., et al.: Abnormal pancreatic glucagon secretion and postprandial hyperglycemia in diabetes mellitus. ].A.M.A., 234:159, 1975. Greve, T., Dayton, A. D., and Anderson, N. V.: Acute pancreatitis with coexistent diabetes mellitus: An experimental study in the dog. Am.]. Vet. Res., 34:939, 1973. Hill, F. W. G.: Malabsorption syndrome in the dog: A study of thirty-eight cases.]. Small Anim. Pract., 13:575, 1972. James, R. C., and Chase, G. R.: Evaluation of some commonly used semiquantitative methods for urinary glucose and ketone determinations. Diabetes, 23:4 74, 1974.

628

WILLIAM

D.

ScHALL, LARRY

M.

CoRNELIUS

17. Kidson, W., Casey, J., Kraegen, E., eta!.: Treatment of severe diabetes mellitus by insulin infusion. Brit. Med. J., 2:687, 1974. 18. Kleeman, C. R., and Liberman, B.: Diabetic acidosis and coma. In Maxwell, M. H., and Kleeman, C. R. (eds.): Clinical Disorders of Fluid and Electrolyte Metabolism. New York, McGraw-Hill, 1972. 19. Levine, R.: Glucagon and the regulation of blood sugar. New Engl.]. Med., 294A94, 1976. 20. Ling, G. V., Lowenstine, L. J., Pulley, L. T., eta!.: Diabetes mellitus in dogs: A review of initial evaluation, immediate and long-term management, and outcome. J.A.V.M.A., 170:521, 1977. 21. Lorenz, M. D., and Cornelius, L. M.: Laboratory diagnosis of endocrinological disease. VET. CuN. NoRTH AM., 6:687, 1976. 22. Mahaffey, M. B., and Anderson, N. V.: Effect of staphylococcal alpha toxin pancreatitis on glucose tolerance in the dog. Am. J. Vet. Res., 37:947, 1976. 23. Matz, R.: Diabetic coma. N.Y. StateJ. Med., 74:642,1974. 24. Molnar, G. D., and Service, F. J.: Low-dosage continuous insulin infusion for diabetic coma. Ann. Intern. Med., 81:853, 1974. 25. Pairent, F. W., Trapnell, J. E., Howard, J. M.: The treatment of pancreatic exocrine insufficiency. Ill. The effects of pancreatic ductal ligation and oral pancreatic enzyme supplements on fecal lipid excretion in the dog. Ann. Surg., 170:737, 1969. 26. Petty, C., and Cunningham, N. L.: Insulin absorption by glass infusion bottles, polyvinylchloride infusion containers, and intravenous tubing. Anesthesiology, 40:400, 1974. 27. Posner, J. B., and Plum, F.: Spinal-fluid pH and neurologic symptoms in systemic acidosis. New Engl. J. Med., 277:605, 1967. 28. Sakurai, H., Dobbs, R., and Unger, R. H.: Somatostatin-induced changes in insulin and glucagon secretion in normal and diabetic dogs. J. Clin. Invest., 54: 1395, 1974. 29. Schaer, M., Scott, R., Wilkins, R., et a!.: H yperosmolar syndrome in the non- ketoacidotic diabetic dog. J.A.A.H.A., 10:357, 1974. 30. Schaer, M.: A clinical survey of thirty cats with diabetes mellitus. J.A.A.H.A., 13:23, 1977. 31. Scherston, B., Kuhl, C., Hollender, A., eta!.: Blood glucose measurements with Dextrostix and new reflectance meter. Brit. Med. ]., 3:384, 1974. 32. Shaw, C. E., Hurwitz, G. E., Schmukler, M., eta!.: A clinical and laboratory study of insulin dosage in diabetic acidosis: Comparison with small and large doses. Diabetes, 11:23, 1962. 33. Unger, R. H., and Orci, L.: The essential role of glucagon in the pathogenesis of diabetes mellitus. Lancet, I: 14, 1975. 34. Wilson, J. W., and Hulse, D. A.: Surgical correction of islet cell adenocarcinoma in a dog. J. A.V.M.A., 164:603, 1974. Department of Small Animal Surgery and Medicine Michigan State University College of Veterinary Medicine East Lansing, Michigan 48824