What to Feed Diabetic Dogs

2. What to Feed Diabetic Dogs

Evidence-based Approach

Recommendations for feeding diabetic dogs should ideally be based on evidence provided by results of randomized, controlled clinical trials that clearly document significant clinical value of the test diet. Whenever this is lacking, clinicians must assess the best evidence that is available and interpret this in the light of expert clinical experience and knowledge of current pathophysiological concepts.

To assist this process, evidence in the following review has been ranked into categories (Table 2):

1. Randomized, controlled, clinical trials in diabetic dogs
2. Other clinical trials in diabetic dogs
3. Randomized, controlled, clinical trials in non-diabetic dogs
4. Expert opinion, clinical experience, and pathophysiological rationale

General Goals of Nutritional Therapy for Diabetic Dogs

Evidence Based on Expert Opinion, Clinical Experience, and Pathophysiological Rationale

The diet of diabetic dogs should provide adequate calories to achieve and maintain optimal body condition. Dogs with poorly controlled diabetes have a decreased ability to metabolize the nutrients absorbed from their gastrointestinal tract and lose glucose in their urine and so may require more calories for maintenance than healthy dogs. The diet fed should be nutritionally balanced and needs to be palatable so that food intake is predictable. Meals should ideally be timed so that maximal exogenous insulin activity occurs during the postprandial period (Church, 1982). Because the daily insulin-dosing regimen tends to be fixed for diabetic dogs, it is also important that a predictable glycemic response is achieved following each meal. Consequently, every meal should contain roughly the same ingredients and calorie content, and should be fed at the same time each day. The owners of diabetic dogs should be aware that a consistent insulin-dosing and feeding routine is optimal.

Dietary Fiber and Canine Diabetes

Total Dietary Fiber

Evidence Based on Various Clinical Trials in Diabetic Dogs

Some studies in diabetic dogs have indicated that high-fiber diets might be associated with improved glycemic control. However, these studies have compared high-fiber (56 - 73 g/1000kcal and 15%DM) with lower-fiber (16 - 27 g/1000 kcal) diets without including comparison with a control diet formulated for typical canine adult maintenance. Thus, there has not been a clear demonstration of clinical benefit for diabetic dogs fed a high-fiber formulation compared with feeding a typical adult maintenance diet.

Additionally, low-fiber diets typically contain increased dietary starch content, which might be a confounding factor when comparing the glycemic responses of diabetic dogs to high- and low-fiber diets. Regardless of the composition of the high-fiber diet or the length of time over which the diabetic dogs were monitored, no significant difference in daily insulin requirement (Nelson et al., 1991; Graham et al., 1994; Nelson et al., 1998, Nelson et al., 2000; Kimmel et al., 2000; Graham et al., 2002) or fasting triglyceride levels (Nelson et al., 1991, Nelson et al., 1998; Graham et al., 2002) between groups of diabetic dogs fed low-fiber and high-fiber diets has been found.

Importantly, there seems to be marked variation between the responses of individual diabetic dogs to dietary fiber. In one study (Nelson et al., 1998), significant improvement of all indices of glycemic control, including lowered daily insulin requirement, was seen in 9 of 11 dogs when they were fed a high-fiber diet (64.4 g/1000kcal). The remaining 2 dogs were found to have improved glycemic control on the lower-fiber diet (27.0 g/1000kcal or 11% in 4000 kcal/kg of food).

In another study of 12 diabetic dogs (Nelson et al., 2000), glycemic control was best in 6 dogs when fed a soy-based, moderate-fiber diet (total dietary fiber 8% DMB), in 4 dogs when fed a cellulose-based, high-fiber diet (total dietary fiber 16% DMB), in 1 dog when fed a cellulose-based, moderate-fiber diet (total dietary fiber 8% DMB), and glycemic response to diet could not be ranked in the remaining dog. A similar situation exists for people because high-fiber diets do not have a uniform effect in all diabetic subjects (EASD, 1988). This might be partly due to the side effects that are sometimes associated with high-fiber diets, which include poor palatability, poor weight gain, poor hair coat, vomiting, voluminous feces, flatulence, diarrhea, and constipation. Individual tolerance to dietary fiber is dependent on a large number of factors, including the quality and type of the fiber.

Evidence Based on a Randomized, Controlled, Clinical Trial in Diabetic Dogs

A randomized, controlled, trial was performed to assess the influence of canned, high-fiber, mode-rate-starch diets on insulin requirement and glycemic control in dogs with stabilized diabetes (Fleeman & Rand, 2003). The two trial diets had high-fiber (50 g/1000kcal) and moderate-starch (26% ME) content, but varied in fat content (31% ME and 48% ME). The control diet was a commercial dog food formulated for adult maintenance with moderate-fiber (35 g/1000kcal), low-starch (2.3% ME), and higher fat (61% ME) content.

Diabetic control evaluated every 2 weeks included history, physical examination, and 2-hourly blood glucose measurements over 12 hours. Insulin dose was adjusted based on standardized criteria to maintain control of glycemia. At the end of each 2 month feeding period, glycemic control was evaluated by plasma fructosamine, glycosylated hemoglobin, and 48 hour serial blood glucose measurements. No significant differences in insulin requirement or glycemic response among diets were found. It was concluded that, for stable diabetic dogs, high-fiber, moderate-starch diets offer no significant advantage for insulin requirement or glycemic control compared with a commercial diet formulated for adult maintenance with moderate-fiber and low-starch content.

Different Types of Dietary Fiber

Evidence Based on Pathophysiological Rationale

Soluble Fiber 

Dietary fiber can be characterized by degree of solubility, which is a reflection of its properties in an aqueous media. Soluble fiber, as provided by guar gum and psyllium, has great water-holding capacity, and forms a viscous solution in the intestine.

Dogs fed diets with increased viscosity might have more rapid postprandial glucose absorption, resulting in higher total postprandial glucose absorption and are more likely to develop secretory diarrhea than dogs fed diets with lower viscosity (Nelson & Sunvold, 1998b).

This suggests that only diets with an intermediate viscosity (solubility) level might be associated with a delay in gastrointestinal transit time and optimal glucose homeostasis in dogs.

Soluble fiber, with the exception of psyllium, is usually also fermentable fiber.

Psyllium grains. The outer husk is high in non-fermentable mucilage that is soluble in water. (©Royal Canin Laboratory).

Fermentable Fiber 

Dietary fiber can be characterized by degree of fermentability, as well as solubility. Fermentable fiber is readily degraded by colonic microflora in dogs to produce short-chain fatty acids that are absorbed across the intestinal mucosa.

Fermentable dietary fiber is associated with increased intestinal glucose transport capacity, increased glucagon-like-peptide-1, and increased insulin secretion in non-diabetic dogs (Massimino et al., 1998). The overall effect is a significant reduction of the area under the blood glucose concentration versus time curve during oral glucose tolerance testing. As diabetic dogs lack the capacity to increase insulin secretion and match increased intestinal glucose transport, it needs to be investigated whether they benefit from diets containing high levels of fermentable fiber or whether these diets may actually contribute to glucose intolerance.

Insoluble, Non-fermentable Fiber 

Dogs cannot digest the insoluble fiber component of their diet and it is excreted in the feces. In contrast to soluble fiber, insoluble fiber such as purified cellulose seems to exert relatively little physiological effect in the canine gut and can be tolerated in fairly high dietary levels (Bauer & Maskell, 1995).

Evidence Based on a Randomized, Controlled, Trial in Non-diabetic Dogs

A randomized controlled evaluation in non-diabetic dogs of the effects of diets containing different fiber types (highly-soluble, highly-fermentable guar gum, poorly-soluble, poorly-fermentable cellulose, and mixed soluble-insoluble, moderately-fermentable sugar beet pulp fiber) at three different dietary concentrations has helped to clarify some of the issues relating to the putative glucoregulatory effects of dietary fiber in dogs (Hoenig et al., 2001) (Figure 5). The different test diets were obtained by substituting 3.5% DM of cornstarch in the control diet with the fiber sources mentioned above. The total dietary fiber level varied between 4.9% and 17.2% DM (Hoenig et al., 2001).

Figure 5. A mixture of beet pulp and cellulose (bran).  

Compared with the control diets (total dietary fiber 3.5% and 4.4% DM), there were no significant differences in physical findings, serum glucose and insulin concentrations during oral glucose tolerance testing, serum triglyceride concentrations, or cholesterol content of HDL, LDL, and VLDL associated with feeding any of the fiber-modified diets. The only significant findings were that total serum cholesterol concentrations were lower in dogs fed sugar beet fiber and higher in dogs fed cellulose fiber, compared with control diets. Although it was not objectively measured, it was noted that the dogs' coat hairs seemed to become dull and lusterless when they consumed the fiber-modified diets.

The authors proposed that this might have been due to an inhibitory effect of fiber on the absorption of minerals and vitamins.

Evidence Based on Various Clinical Trials in Diabetic Dogs

When dogs were fed a single meal containing added soluble fiber or added insoluble fiber, a greater reduction of postprandial hyperglycemia was seen with the meal containing soluble fiber (Blaxter et al., 1990) although the dietary fiber composition of the diets was not reported and were probably not comparable (Davis, 1990).

When comparisons were made following long-term feeding for 1 or 2 months of diets high in soluble fiber or insoluble fiber (34 g/1000 kcal soluble fiber versus 60 g/1000 kcal insoluble fiber) (Nelson et al., 1991); 10 g/1000 kcal soluble fiber versus 73 g/1000 kcal insoluble fiber (Kimmel et al., 2000), a tendency for improved glycemic control and fewer side effects was seen with the diets containing increased insoluble fiber. In particular, significantly lower glycosylated hemoglobin (Nelson et al., 1991) or fructosamine (Kimmel et al., 2000) levels were recorded. The current evidence regarding dietary fiber and canine diabetes mellitus is summarized in Table 3.

Dietary Carbohydrate and Canine Diabetes

Total Dietary Carbohydrate

Evidence Based on Randomized Trials in Non-diabetic Dogs

The amount of starch in the diet has been shown to be the major determinant of the postprandial glycemic response of healthy dogs across 15 typical commercial dog foods (dietary starch 0.4 - 52.7% DMB), regardless of the carbohydrate source or type, or of the composition profile of other macronutrients (Nguyen et al., 1998b). Although similar studies have not been performed in diabetic dogs, there is very good evidence in diabetic people for a strong association between the insulin dosage requirement and the carbohydrate content of the meal, regardless of the glycemic index (Figure 6), the carbohydrate source or type, or the composition profile of other macronutrients (Franz et al., 2002a). The same might be true for diabetic dogs.

Figure 6. What is the glycemic index ? In man, GI (glycemic index) does not necessarily represent a practical guide for evaluating foods because data can be in conflict depending on the composition of the meal, the processing method, cooking, etc. Answers can also vary amongst individuals. In animals, results are more reliable because the diet can be better controlled.  

Different Types of Dietary Carbohydrate

Evidence Based on Physiological Rationale

The postprandial glycemic response to dietary carbohydrate might be potentially influenced by the type of carbohydrate and by the way it has been processed. Digestion of dietary carbohydrate occurs in the small intestine of dogs and results in the breakdown of starch to glucose, fructose, and galactose. The postprandial glycemic response is directly dependent on the absorption of glucose, because fructose and galactose require hepatic metabolism for conversion to glucose. Thus, the type of starch contained in the dietary carbohydrate fed might influence the postprandial glycemic response. Carbohydrate sources that predominantly breakdown to glucose during digestion are likely to result in the greatest postprandial glycemic response.

Studies assessing the digestibility in dogs of different dietary carbohydrate substrates (Murray et al., 1999; Bednar et al., 2000; Twomey et al., 2002), have found that the processing method as well as the carbohydrate source significantly influences digestibility (Bednar et al., 2000). For example, barley flour is approximately five times more digestible in dogs compared with barley grain, while rice flour is almost ten times more digestible than white rice grain (Bednar et al., 2000). When commercial dog foods are formulated, the dietary starch is usually in the form of flours prepared using a combination of roller milling, sieving, and steam cooking (Murray et al., 1999). The extrusion process then tends to gelatinize the starch and make it even more digestible (Camire, 1998), so that starch digestibility is essentially 100% for most carbohydrate sources included in commercial dry dog foods (Murray et al., 1999; Twomey et al., 2002). There is some evidence that the gelling agents used in canned commercial dog foods may similarly increase digestibility (Karr-Lilienthal et al., 2002). Thus, for most commercial dog foods, processing effects likely have minimal influence on the post-grains (in this case rice) is lower prandial glycemic response is likely the dietary carbohydrate source.

The digestibility of whole cereal grains (in this case rice) is lower than that of the same cereal ground into meal.

Evidence Based on a Randomized Trial in Non-diabetic Dogs

Little is known about the glycemic responses of diabetic dogs to different sources of dietary carbohydrate. However, a study in non-diabetic dogs that examined the postprandial effects of five diets with equivalent starch content (30% DMB) from different cereal sources found marked differences in the glucose and insulin responses (Sunvold & Bouchard, 1998; Bouchard & Sunvold, 2001). The rice-based diet resulted in significantly higher postprandial glucose and insulin responses. Sorghum generally caused the lowest postprandial glucose response while barley produced the lowest insulin response. These findings form an interesting basis for future study on the effects of diets containing sorghum in diabetic dogs, but more work is required before specific recommendations can be made. Caution is required when extrapolating the results of dietary carbohydrate studies in non-diabetic dogs to clinical recommendations for diabetic dogs. This is because all diabetic dogs require exogenous insulin therapy, which has an overwhelming effect on carbohydrate metabolism and the postprandial glycemic response. It is also worth noting that studies in people have found a marked variability in the glycemic response to different types of barley (Liljeberg et al., 1996) and rice (Jarvi et al., 1995). The same is likely true for dogs.

The current evidence regarding dietary carbohydrate and canine diabetes mellitus is summarized in Table 4.

Dietary Fat and Canine Diabetes

Evidence Based on Expert Opinion, Clinical Experience, and Pathophysiological Rationale

Altered lipid metabolism occurs with insulin deficiency in dogs, yet there are minimal published data on the influence of dietary fat on diabetic dogs. In human patients, the lipid disorders that occur in association with diabetes are atherogenic and predispose to coronary artery disease (Stamler et al., 1993). Restricted-fat diets reduce cardiovascular morbidity and mortality in diabetic people. Although atherosclerosis and coronary artery disease are not usually a clinical concern in diabetic dogs, atherosclerosis does occur in association with spontaneous canine diabetes (Sottiaux, 1999; Hess et al., 2003). Perhaps of greater clinical relevance is that diabetes secondary to exocrine pancreatic disease appears to be common in dogs, and the diabetic state might also be a risk factor for pancreatitis. High-fat diets and hypertriglyceridemia have been proposed as possible inciting causes of canine pancreatitis (Simpson, 1993; Williams, 1994). Low-fat diets (for example fat < 20% ME) are recommended for dogs with chronic pancreatitis. As it can be difficult to identify those diabetic dogs with subclinical pancreatitis (Wiberg et al., 1999), it might be prudent to consider feeding a restricted-fat diet (for example fat <30% ME) to all diabetic dogs. This might have the added benefit of improving insulin sensitivity in animals with insulin resistance-associated diabetes and reducing the risk of overt diabetes in bitches during diestrus. However, greater levels of energy restriction might lead to undesirable weight loss.

Evidence Based on a Randomized, Controlled, Clinical Trial in Diabetic Dogs

The same randomized, controlled trial that assessed the influence of canned, high-fiber, moderate-starch diets on insulin requirement and glycemic control of dogs with stabilized diabetes also assessed the influence of dietary fat (Fleeman & Rand, 2003). Different amounts of dietary fat in the high-fiber (50 g/1000 kcal), moderate starch (26 % ME) diets had no significant influence on insulin requirement or glycemic control of the dogs. Lower dietary fat content (31% ME compared with 48% ME) was associated with significantly improved lipid profiles. The low fat, high fiber, moderate starch diet resulted in significantly lower mean total cholesterol concentration compared with either of the other diets, and significantly lower mean glycerol and free fatty acids than the commercial diet. It is unknown whether any health benefits for dogs might be attributed to these improvements in the lipid profile. Significant weight loss occurred when the dogs were fed the low-fat, high-fiber, moderate-starch diet, whereas maintenance of weight was achieved with both of the other diets. It was concluded that diets with lower fat content may result in improved lipid profiles in diabetic dogs, but might contribute to undesirable weight loss. Therefore, restricted-fat diets should not routinely be recommended for diabetic dogs with thin body condition.

The current evidence regarding dietary fat and canine diabetes mellitus is summarized in Table 5.

Dietary Protein and Canine Diabetes

Evidence Based on Pathophysiological Rationale

The optimal dietary protein for diabetic dogs has not been determined and it is rational that recommendations would be no different than for non-diabetic dogs. As restriction of dietary carbohydrate might reduce postprandial hyperglycemia in diabetic dogs and dietary fat restriction might be beneficial if there is concurrent pancreatitis, there will be a tendency for suitable diets to have higher protein levels (>30% ME).

Microalbuminuria and proteinuria do occur in diabetic dogs (Struble et al., 1998) and lower dietary protein intake may be indicated in diabetic dogs with microalbuminuria.

Dietary L-carnitine and Canine Diabetes

Evidence Based on Pathophysiological Rationale

L-Carnitine is a conditionally essential, vitamin-like nutrient that plays a pivotal role in fatty acid metabolism. Supplemental L-Carnitinesuppresses acidosis and ketogenesis during starvation in dogs (Rodriguez et al., 1986). L-Carnitine supplementation at 50 ppm of diets fed to dogs enhances energy conversion from fatty acid oxidation and protects muscles from catabolism during weight loss (Gross et al., 1998; Sunvold et al., 1999; Center, 2001). Dogs with poorly controlled diabetes experience weight loss, altered fat metabolism, ketogenesis, and hepatic changes, and so are likely to benefit from dietary L-carnitine supplementation. The majority of diabetic dogs are middle-aged and older and can be expected to already have reduced lean body mass (Kealy et al., 2002) before the onset of diabetes-associated weight loss. Consequently, it is important to consider any dietary intervention, such as L-carnitine supplementation, that promotes maintenance of lean body mass in these animals.

Dietary Chromium and Canine Diabetes

Evidence Based on Pathophysiological Rationale and a Controlled Clinical Trial in Diabetic Dogs

Chromium tripicolinate is a dietary mineral supplement that has been shown to increase the clearance rate of glucose from the blood by approximately 10% in healthy dogs (Spears et al., 1998). However this potential benefit is only possible if there is chromium deficiency because chromium is a nutrient, not a drug. Thus, supplementation may only result in benefits if the individual is deficient or marginally deficient in chromium.

It is now clear that dietary chromium levels of people in industrialized countries are sub-optimal (Anderson, 1998). Similar information is not available for dogs and further studies are warranted to try and establish the minimum recommended dietary chromium intake for healthy dogs.

Chromium is thought to potentiate insulin's ability to store glucose and would theoretically be useful in dogs with insulin resistance or as an adjunct to exogenous insulin therapy. It is also possible that inadequate dietary intake of chromium by dogs might increase their risk of developing diabetes. It has been postulated that some insulin-dependent diabetic people might lose their ability to convert inorganic chromium to the biologically active form and might actually need to consume foods that contain active forms of chromium (Anderson, 1992). At this stage, there is little information available on the effects of chromium supplementation in human patients requiring insulin therapy (Ravina et al., 1995; Fox et al., 1998). Supplementation with chromium picolinate capsules has not been found to improve glycemic control in insulin-treated dogs (Schachter et al., 2001). The influence of chromium supplementation on bitches with diestrus-induced insulin resistance is unknown.

Dietary chromium supplements usually contain low molecular weight chromium salts such as trivalent chromium [Cr(III)], which has a large safety margin but can be toxic at very high doses (Jeejeebhoy, 1999). In contrast, oral hexavalent chromium [Cr(VI)] appears to be 10 - 100 times more toxic than trivalent chromium compounds and is an unsuitable dietary supplement (Katz & Salam, 1993).

Summary of Dietary Recommendations for Canine Diabetics

The American Diabetes Association uses a grading system to rank the scientific principles of their nutritional recommendations.
-The highest ranking, Grade A, is assigned when there is supportive evidence from multiple, well-conducted studies
- Grade B is an intermediate rating
- Grade C is a lower ranking
- Grade E represents recommendations based on expert consensus.

If this grading system is used to rank the scientific basis of the nutritional recommendations for canine diabetes, current evidence can be summarized in the following fashion.

Grade B Evidence
- Controlled evaluation in non-diabetic dogs of diets with different amounts and types of fiber indicate that increased fiber intake has no significant influence on glucose homeostasis, compared with typical diets formulated for canine adult maintenance.
- Several studies in diabetic dogs indicate that high-fiber diets, compared with low-fiber diets, might be associated with improved glycemic control. However, randomized, controlled comparison identified no measurable benefit for insulin requirement or glycemic control in diabetic dogs, compared with a conventional, moderate-fiber diet formulated for adult maintenance (Grade C evidence).
- There seems to be marked variation between the responses of individual diabetic dogs to dietary fiber.
- High-fiber diets do not significantly improve hypertriglyceridemia in diabetic dogs but might lower serum cholesterol concentrations.
- Supplementation with chromium capsules has not been found to improve glycemic control in insulin-treated dogs.

Grade C Evidence
- When lower-fiber diets are fed to diabetic dogs, a blend of soluble and insoluble fibers (such as soy fiber or beet pulp) might be preferable to insoluble fiber alone.
- Comparison in non-diabetic dogs found that a rice-based diet resulted in significantly higher postprandial glucose and insulin responses, while a sorghum-based diet caused reduced glucose responses, and barley produced lower insulin responses.
- Diabetic dogs might benefit from dietary L-carnitine supplementation.
- Diets with lower fat content might result in improved lipid profiles in diabetic dogs, but might also contribute to undesirable weight loss.

Grade E Evidence
- The diet fed to diabetic dogs should be palatable so that food intake is predictable
- The diet fed to diabetic dogs should be nutritionally balanced.
- The nutritional requirements of any concurrent disease may need to take precedence over the dietary therapy for diabetes.
- As a regimen of fixed daily insulin dosages is typically used to manage diabetic dogs, it is rational to provide a consistent amount of carbohydrate in the meals fed each day.
- The optimal dietary protein for diabetic dogs has not been determined. Lower dietary protein might be indicated only in diabetic dogs with microalbuminuria or proteinuria.