Treatment of feline diabetes mellitus using an α-glucosidase inhibitor and a low-carbohydrate diet

Abstract

The purpose of this study was to determine the effect of an α-glucosidase inhibitor (acarbose), combined with a low-carbohydrate diet on the treatment of naturally occurring diabetes mellitus in cats. Eighteen client-owned cats with naturally occurring diabetes mellitus were entered into the study. Dual-energy X-ray absorptiometry (DEXA) was performed prior to and 4 months after feeding the diet to determine total body composition, including lean body mass (LBM) and percent body fat. Each cat was fed a commercially available low-carbohydrate canned feline diet and received 12.5 mg/cat acarbose orally every 12 h with meals. All cats received subcutaneous insulin therapy except one cat in the study group that received glipizide (5 mg BID PO). Monthly serum glucose and fructosamine concentrations were obtained, and were used to adjust insulin doses based on individual cat's requirements. Patients were later classified as responders (insulin was discontinued, n=11) and non-responders (continued to require insulin or glipizide, n=7). Responders were initially obese (<28% body fat) and non-responders had significantly less body fat than responders (<28% body fat). Serum fructosamine and glucose concentrations decreased significantly in both responder and non-responder groups over the course of 4 months of therapy. Better results were observed in responder cats, for which exogenousinsulin therapy was discontinued, glycemic parameters improved, and body fat decreased. In non-responders, median insulin requirements decreased and glycemic parameters improved significantly, despite continued insulin dependence. The use a low-carbohydrate diet with acarbose was an effective means of decreasing exogenous insulin dependence and improving glycemiccontrol in a series of client-owned cats with naturally occurring diabetes mellitus.

Introduction

The incidence of diabetes mellitus in the domestic cat population has been increasing from one in 1500 cats in the 1960s and 1970s (Meier 1960, Shaer 1977) to one in 250 cats in the 1990s (Panciera et al 1990). Treatment of diabetes in cats is often challenging due to waxing and waning blood glucose levels and variable degrees of insulin resistance (Goossens et al 1998). Recent studies have demonstrated that the pathogenesis of diabetes mellitus in most domestic cats is similar to Type-2 or non-insulin-dependent diabetes mellitus in humans (Johnson et al 1986, 1989, Lutz and Rand 1993, Lutz et al 1994, O'Brien et al 1985).

Diets high in complex carbohydrates, such as fiber, have been recommended for human Type-2 diabetic patients. A recent study documented a significant decrease in mean pre-prandial and 12-h mean serum glucose concentrations in diabetic cats fed a diet high in insoluble fiber (HF) compared with a low fiber diet (LF); however, no change in daily insulin requirements or bloodglycosylated hemoglobin concentrations were observed (Nelson et al 2000). Human beings and cats with Type-2 diabetes mellitus have been shown to have improved glycemic control and improved nitrogen turnover during weight loss when a low-energy, high-protein diet was combined with oral hypoglycemic therapy or insulin (Anderson et al 2001, Gougeon et al 2000). The α-glucosidase inhibitors, such as acarbose, impair glucose absorption from the intestine by decreasing starch digestion, and hence, glucose production from food sources (Kahn and Shechteer 1990). The activity of acarbose, when combined with insulin and diet, can improveglycemic control in dogs and cats (Greco 1998, Robertson et al 1999).

Measurement of body composition using dual-energy X-ray absorptiometry (DEXA), a quantitative measure of body composition, has been used to assess carbohydrate metabolism (Lukaski 1993, Massimino et al 1999). Previous studies of body composition in the cat have shown that a higher percentage of body fat is associated with greater glucose intolerance (Massimino et al 2000). Obesity has been demonstrated to be a predisposing factor to insensitivity to insulin and the development of Type-2 diabetes mellitus in both cats and human beings.

The purpose of this study was to determine whether the use of a low-carbohydrate diet, with or without acarbose, was effective in improving glycemic control, reducing insulin requirements, and changing body composition in a series of cases of client-owned cats with naturally occurring diabetes mellitus.

Materials and methods

Patient selection

Following informed client consent and approval by the Colorado State University Animal Care and Use Committee, 24 client-owned cats (18 diet/acarbose, six diet only controls) with previously diagnosed diabetes mellitus wereincluded in the study. Twenty cats had previously been treated with insulin for a period ranging from 2 weeks to 3 years. No cat previously treated with insulin was a controlled diabetic upon entry into the study. Four cats had not received previous insulin therapy before entrance into the study. All cats were subjectively overweight based on body condition scores (6–9 of 9), or historically were overweight before clinical weight loss became apparent. Physical examination, baseline serum thyroxine, serum fructosamine concentration, serum glucose concentration, complete blood count, and urinalyses were performed to rule out the presence of illnesses other than diabetes mellitus. Cats with evidence of concurrent diseases other than diabetes mellitus (ie urinary tract infection, hyperthyroidism, pancreatitis, etc) were excluded from the study. No cat was ketotic at the time of entry into the study.

Dual-energy X-ray absorptiometry

Each patient had an intravenous catheter placed and was anesthetized with intravenous Propofol (Abbott Laboratories, North Chicago, IL, 4–7 mg/kg IV bolus, then continuous infusion to effect) for each DEXA scan. Supplemental oxygen was administered and routine anesthetic monitoring was performed. A DEXA scan (Hologic QD 1000/W with software version 5.71P, Bedford, MA) was obtained from each patient to obtain exact measurements of total body composition. Cats were positioned in sternal recumbency with the hind limbs extended caudally, and DEXA measurements were performed using a whole body scanner operated in single-beam mode, as described previously (Grier et al 1996). Calibration of the unit was verified by scanning a calibration phantom. Commercially available pediatric software was used to analyze the scans. Body mass (lean and fat) was obtained, and the percent body fat was calculated. Scans took approximately 10 min and were performed in duplicate if any gross movement was detected. In one cat, five consecutive DEXA scans were performed in a 1-h time period, repositioning in between each scan, to determine coefficient of variation for this technique. The DEXA scans were repeated after 4 months of therapy.

Treatment

Each cat received insulin (Lente, Humulin-L, Lilly, Indianapolis, IN, n=18); or PZI (Protamine Zinc Insulin, Blue Ridge Pharmaceuticals, Greensboro, NC, n=6) therapy as recommended by the attending veterinarian. In addition to twice daily insulin injections (12 h apart), each cat was fed a commercially available low-carbohydrate canned diet (Hill's Science Diet Feline Growth canned, Hill's Pet Food Inc., Topeka, KS) consisting of approximately 49% protein, 36.2% fat, and 6.9% carbohydrate on a dry matter basis. Each patient's total daily energy requirements were estimated using the formula [30 (body weight_kg) + 70] and provided in two meals fed 12 h apart. No other treats or any dry food were allowed. In addition to the canned diet, each of the study cats received 12.5 mg acarbose(Precose^®, Bayer, West Haven, CT) per os every 12 h with the meal.

Glycemic parameters

Baseline serum fructosamine and glucose concentrations were analyzed for each patient at the time of entry into the study. Post-prandial serum glucose (obtained 2–4 h after the meal) and fructosamine analyses were obtained at monthly intervals for a period of 4 months. Changes in daily exogenous insulin administration were made based on serum fructosamine concentrations and resolutions of clinical signs within the past month (Thorensen and Bredal 1996).

Classification and statistics

Non-responders were defined as continuing to require exogenous insulin for glycemic control, and exhibiting elevations in serum fructosamine <400 μmol/l. Alternatively, responders did not continue to require exogenous insulin for maintenance of normal serum fructosamine concentration (<400 μmol/l). After evaluating the data, within each group (responders versus non-responders), patient body composition was classified as below normal (<28% body fat) or obese (<28% body fat) based on previously documented values (Massimino et al 1999).

A modified Kolmogorov–Smirnov test was used to assess normality of distribution of data, and a Bartlett's test of homogeneity to assess equivalence of variances among groups to determine the distribution characteristics and whether the actual data could be used for parametric analysis. All parameters were normally distributed, and were compared between groups by an analysis of variance for treatment effects, with repeated measures for time. Fisher's least significant difference test was used to identify individual group/time differences. Values of P<0.05 were considered to be significant. All results are presented as mean±standard deviation.

Results

Thirteen male and five female cats were included in the acarbose/diet group. The control group consisted of five males and one female. Median age was 10 years for both groups. All cats had a history of obesity. Physical examination parameters such as hair coat cleanliness and the appearance of seborrhea sicca improved in all cats with therapy. Four cats in the acarbose/diet group had diabetic neuropathy as evidenced by a plantigrade rear limb stance; all cases of neuropathy resolved with therapy. All but four cats, which were newly diagnosed diabetics, had received previous insulin therapy. Poorly regulated diabetic cats were fed a variety of dry feline diets prior to entry into the study. No cat was being fed exclusively a canned diet at the time of entry into the study.

In the acarbose/diet group, 11 cats were classified as responders, and seven cats as non-responders. In the diet control group, four cats were classified as responders and two as non-responders. The total daily median insulin dose in non-responders decreased from 10 U/cat/day (5 units/dose) to 2 U/cat/day (1 unit/dose) in both the diet and control groups.

No significant difference in serum fructosamine or glucose concentrations was observed between responders and non-responders before or after treatment in either the study or control groups. However, serum glucose (Δ175±195 mg/dl) and fructosamine concentrations (Δ165±152 μmol/l) decreased significantly (P<0.01) in all cats (responder and non-responder in study and control groups) after therapy (Fig. 1, Table 1) All responders had improved glycemic parameters, as evidenced by decreased serum glucose values (357±173 to 122±39 mg/dl) and decreased serum fructosamine concentrations (488±231 to 284±6787 μmol/l) despite discontinuation of exogenous insulin therapy. Non-responders also displayed improved glycemic control, with decrease in serum glucose (426±109 to 344±123 mg/dl) and serum fructosamine (638±158 to 529±100 μmol/l) concentrations, but with continued insulin therapy.

Table 1.

Age, sex, pre- and post-percentage body fat, serum fructosamine, serum blood glucose (BG), insulin dose, and diet prior to study

R or N	Age (years)/sex	Pre-weight (kg)	Post-weight (kg)	Pre-percentfat	Post-percentfat	Pre-fructosamine (μmol/l)	Post-fructosamine (μmol/l)	Pre-BG (mg/dl)	Post-BG (mg/dl)	Pre-insulindose (units)	Post-insulindose (units)	Diet prior to study
R	9/MN	6.7	7.6	42.4	36.7	538	298	408	129	L, 2 U∗	0	Dry, fiber
R	10/MN	4.8	5.4	28.1	27.2	272	245	436	145	L, 3 U∗	0	Dry, store
R	12/FS	6.6	6.1	48.6	44.8	261	209	41	162	L, 14 U	0	Dry, fiber
R	12/FS	5.3	4.3	42	30.1	122	272	85	146	L, 8 U	0	Dry, fiber
R	10/MN	7.1	7.1	29.1	28	891	351	330	126	L, 4 U	0	Dry, maint
R	16/FS	4.9	5.2	37.6	33.9	290	157	646	117	L, 6 U	0	Dry, lite
R	13/FS	5.6	5.3	45.4	42.9	522	337	426	138	PZI, 2 U∗	0	Dry, lite
R	10/MN	7.9	8.6	45.5	40	505	303	445	176	L, 4 U	0	Dry, lite
R	6/FS	5.2	6.4	39.6	33.3	633	350	266	89	L, 15 U	0	Dry, fiber
R	12/MN	4.5	4.5	38.7	31.2	736	237	469	48	L, 3 U	0	Dry, fiber
R	10/MN	4.2	5.0	28	28	604	369	379	74	L, 4 U	0	Dry, maint
NR	10/MN	5.5	6.4	15.6	17.7	569	512	498	398	PZI, 3 U	PZI, 1 U	Dry, fiber
NR	10/MN	4.4	5.9	18	26.6	578	424	554	519	L, 4 U	PZI, 2 U	Dry, store
NR	10/MN	5.5	6.4	11.6	24.5	950	646	426	138	L, 8 U	L, 4 U	Dry, senior
NR	9/MN	4.4	5.2	18	25.6	604	689	420	436	L, 4 U	L, 4 U	Dry, kitten
NR	13/MN	6.4	5.8	11.5	23.4	634	462	414	322	PZI, 4 U	PZI, 2 U	Dry maint
NR	8/MN	6.3	6.9	26.5	31	437	460	206	330	0	Glipizide	Dry store
NR	8/MN	4.6	5.0	19	19.7	696	515	460	265	L, 4 U	L, 6 U	Dry, ad lib
CR	9/MN	8.9	9.0	42.9	42.9	730	324	514	121	L, 6 U	0	Dry, kitten
CR	5/MN	6.4	7.1	35.3	31.2	653	324	467	285	L, 6 U	0	Dry, fiber
CR	15/FS	5.1	5.4	39.6	38.5	644	291	556	144	PZI, 4 U	0	Dry, store
CR	9/MN	5.6	6.7	21.9	20.9	657	273	441	139	L, 2 U	0	Dry, store
CNR	10/MN	7.0	7.2	20.9	24.9	653	514	72	192	L, 14 U	L, 8 U	Dry, fiber
CNR	9/MN	4.7	4.8	14.1	16.1	653	432	500	192	PZI, 4 U	PZI, 4 U	Dry, store

(R, responder; NR, non-responder; CR, control responder; CNR, control non-responder; L, Lente; PZI, Protamine Zinc Insulin).

Fig 1.

Box plots depicting the median and range for percentage body fat, fasting blood glucose, and serum fructosamine in 11 responder (<28% body fat) diabetic cats and eight non-responder (<28% body fat) diabetic cats before and after being fed a low-carbohydrate, high-protein diet and acarbose (12.5 mg/cat BID PO) for 16 weeks.

The validated DEXA technique had a coefficient of variation for total body weight, lean body mass (LBM), and percent body fat of 0.3, 2.7, and 3.8%, respectively. Overall, body weight was 5.7±1.2 kg for both groups at the start of the study and 5.9±1.2 kg after 4 months of therapy. No significant difference in LBM (3.85±0.67 kg) was observed between responders (3.71±0.69 kg) and non-responders (4.08±0.62 kg,) at the start of the study. However, LBM (0.26±0.32 kg) was significantly increased in both responders and non-responders in both study and control groups after 4 months of therapy (P=0.0006).

At the onset of the study, responders had a significantly (P<0.0001) higher percent body fat (39±6.7%) compared with non-responders (17.2±5.1%). Following the 4 months of therapy, all responders showed an increase in LBM (271±393 g) and a decrease in body fat (3.51±.75%). All non-responder cats also exhibited an increase in LBM (247±240 g) with therapy. However, the non-responders also showed an increase in body fat (6.3±0.3%). Although all responders showed a decrease in body fat and all non-responders showed an increase in body fat, the change was not statistically significant over the course of therapy.

Discussion

Our population of client-owned cats with naturally acquired diabetes mellitus was similar to that reported by others, in which older, obese male cats were over-represented (Panciera et al 1990). The results of our study demonstrate improved glycemic control in all diabetic cats fed a low-carbohydrate diet alone or in combination with acarbose.

Most of the cats in this study had been previously treated with insulin which supports the observation that optimal response is observed when glucose toxicity is treated with exogenous insulin prior to initiating a change in diet and/or oral hypoglycemic therapy (Struble and Nelson 1997). All cats showed improved glycemic parameters with significant decreases in serum glucose and fructosamine concentrations. However, the best response (discontinuation of insulin therapy) was observed in obese cats in both the study and control groups. This may be due to earlier owner recognition of clinical signs, diagnosis, and initiation of therapy in these patients, before the onset of weight loss and clinical cachexia.

Cachexia is often observed in the later stages of diabetes mellitus. Because the pathogenesis of feline diabetes mellitus includes a degree of endocrine pancreatic exhaustion pursuant to chronic overstimulation of insulin secretion and amyloid deposition (Lutz and Rand 1995, Struble andNelson 1997), it seems plausible that early recognition of the disease can help prevent irreversible exhaustion of pancreatic beta cell insulin secretory capacity. Patients with lower initial body fat (ie non-responders) continued to require exogenous insulin therapy to control signs of diabetes. This may be associated with advanced disease and pancreatic islet cell exhaustion.

The combination of a high dietary carbohydrate load, neutering, decreased exercise, and insulin resistance leads to the development of obesity in this species, and can contribute to the eventual development of Type-2 diabetes mellitus(Biourge et al 1997, Fettman et al 1997, Kirk et al 1993, Link and Rand 1998, Nelson et al 1990). A decrease in dietary carbohydrate load favors insulin secretion by the feline pancreas (Kettlehut et al 1978, Kienzel 1993, Kirk et al 1993, Kitamura et al 1999). This effect was best observed in obese patients who likely had a higher functional pancreatic reserve.

A frequent observation in this series of cases was a worsening of long-term glycemic control with owner non-compliance and addition of small amounts of high-carbohydrate food (mostly dry formulations). This further supports the contention that the benefits observed were associated primarily with a change to the low-carbohydrate diet and not solely by continued use of exogenous insulin in these patients. Since the responder group discontinued insulin therapy, exogenous insulin could not have been the cause of improved diabetic regulation in these cats and theirimprovement could only have been a result of the change in diet. Previous studies have shown that diabetic cats may respond (30% discontinuedinsulin and all cats showed a 50% decrease in insulin requirements) to low dietary carbohydrate therapy alone after treatment of glucose toxicity with exogenous insulin (Anderson et al 2000).

Six client-owned diabetic cats treated with a low-carbohydrate diet alone were also included. We saw the same proportion of cats discontinue insulin on the low-carbohydrate diet and the same changes in body composition. The lack of significant difference between those groups receiving acarbose and those receiving diet alone would suggest that acarbose had a minimal effect on lowering blood glucose and fructosamine concentrations and inducing insulin independence. However, the small numbers of cats in each diet group (with acarbose or without) precludes dismissal of a β statistical error. In other words, a small difference (50% improvement) might be missed without a large number of cats in each group (<50).

The effect of a low-carbohydrate, protein-replete diet on body composition was particularly enlightening. The best response was seen in obese cats exhibiting a decrease in percentage of body fat and an increase in LBM. This effect might have been missed if body weight alone had been measured during the dietary period. However, it became apparent that increases in body weight were associated with an increase in LBM rather than an increase in adipose tissue. Our findings of improved glycemic control and loss of insulin dependence in obese diabetic cats that lost adipose tissue during consumption of a low-carbohydrate diet is also consistent withearlier findings in humans. Human Type-2 diabetics who gain LBM also experience improved glycemic control when consuming low-carbohydrate, high-protein diets (Gougeon et al 2000). Cats that gain weight experience a decrease in insulin responsiveness that is reclaimed upon subsequent loss of body fat (Fettman et al 1998).

In summary, obese cats receiving a low-carbohydrate canned diet showed a loss of insulin dependence and improvement in glycemic control after 4 months of therapy. Cats with low body fat also improved glycemic parameters but continued to require insulin at much lower dosage than prior to the dietary intervention. The addition of acarbose did not seem to affect the number cats that continued to require exogenous insulin or the improvement in glycemic control.