Increased blood insulin levels are associated with an increased risk of pasture-associated laminitis in equids.
Increased blood insulin levels are associated with an increased risk of pasture-associated laminitis in equids.
To determine the relationship between plasma insulin, leptin, and lipid levels, and measures of oxidative stress with adiposity in mature light breed horses.
300 randomly selected light breed horses, aged 4–20 years.
A random sample of horses (140 mares, 151 geldings, and 9 stallions) was drawn from the VMRCVM Equine Field Service practice client list. Evaluations occurred June 15 – August 15, 2006, with all sampling performed between 0600 and 1200 hours. Concentrate feed was withheld for at least 10 hours before sampling. Plasma was analyzed for insulin, glucose, leptin, triglycerides, nonesterified fatty acids, and measures of oxidative stress. Body condition score was determined as the average of 2 independent investigators.
Overconditioned and obese horses had higher plasma insulin (P < .001) and leptin (P < .01) levels than optimally conditioned horses. Obese horses had higher triglyceride levels (P = .006) and lower red blood cell gluthathione peroxidase activities (P = .001) than optimally conditioned horses.
Maintaining horses at a BCS <7 might be important for decreasing the risk of pasture-associated laminitis.
average neck circumference
analysis of variance
body condition group
body condition score
body mass index
neck circumference:height ratio
nonesterified fatty acids
Virginia-Maryland Regional College of Veterinary Medicine
Laminitis, a potentially debilitating disease of equids, is second only to colic as the most common reason for a horse or pony to require veterinary care. Nearly 20% of afflicted animals never fully recover, and many need to be humanely destroyed. For many years, veterinarians have anecdotally observed that horses and ponies have increased laminitis susceptibility while maintained on pasture (pasture-associated laminitis, PAL) if they have excess general or regional adiposity. It is now recognized that many animals with this obesity phenotype that develop PAL are insulin resistant, hyperinsulinemic, or both. Serum insulin levels are higher in ponies with a history of recurrent PAL when compared with nonlaminitic ponies, whereas an insulin concentration >32 mU/L in early spring was predictive of future episodes of PAL in a single pony population over 2 successive years.[2, 3] Furthermore, at an equine hospital specializing in both first opinion and referral cases, hyperinsulinemia was detected in 31 of 36 (86%) horses with laminitis. The recent discovery that prolonged intravenous administration of insulin to healthy ponies and horses induces histologic and clinic signs of laminitis adds further evidence to the association between hyperinsulinemia and laminitis.[5, 6]
There is strong sentiment in the veterinary community that insulin resistance (IR) and hyperinsulinemia (HI) are growing problems in the general equine population. However, there is little information on the prevalence of these conditions in equid populations, nor is there a complete understanding of the relationships between adiposity and IR or HI in these populations. The primary objectives of this study were to investigate the association between plasma insulin and measures of adiposity in a subpopulation of mature horses in Virginia. Secondary objectives were to determine the relationships between age, sex, and exercise level, on levels of insulin, glucose, nonesterified fatty acids (NEFA), triglycerides, leptin, and measures of insulin sensitivity, and oxidative stress.
This study was completed in 60 days (June 15 to August 15, 2006), and was approved by the Institutional Animal Care and Use Committee at Virginia Tech.
Selection of horses, and the collection of health and exercise history has been previously described. Exercise was categorized as none, light, moderate, or intense using the guidelines of the National Research Council. Briefly, a random sample of 300 adult horses (4–20 years old) from a population >1,000 horses was selected from the Virginia-Maryland Regional College of Veterinary Medicine (VMRCVM) Equine Field Service client list. Ponies, miniature horses, mules, donkeys, draft breeds, pregnant and lactating mares, and horses undergoing treatment for health problems were excluded from the study population. All horses were held off concentrate feed for a minimum of 10 hours before sampling. Horses that were stalled overnight were not given more than 2 flakes of hay after 2000 hours the night before sampling and were not turned out to pasture before sampling (restricted access to forage before sampling). Horses that were on pasture since at least 2000 hours the night before sampling or that had a continuous source of hay (ie, round bale hay in a dry lot), or continuous access to both pasture and hay (ie, round bale hay in the pasture) were allowed to continue to consume forage before sampling (continuous access to forage before sampling). Evaluations occurred between 0600 and 1200 hours.
Two independent scorers determined body condition score (BCS), using whole number scores only, and scores were averaged and rounded to the next greater integer. For this study, horses with a BCS of <4 were removed from the analysis (n = 5), horses with a BCS of 4–6 (n = 141) were classified as optimal condition, BCS of 7 (n = 9) as overcondition, and BCS of 8–9 (n = 56) as obese, which determined the body condition groups (BCG). Morphometric measurements were taken on each horse in order to calculate body weight (BW) and body mass index (BMI). Girth circumference, height at the withers, and length from the point of the shoulder to the ipsilateral tuber ischium were measured (cm). BW was estimated from the formula: BW (kg) = [girth (cm)2 × length (cm)]/11,877. BMI was calculated as estimated weight (kg)/height (m)2. Neck circumference (cm) was measured in 3 locations as previously described. These 3 measurements were averaged to determine average neck circumference (ANC). The neck circumference height ratio (NCHR) was calculated as [ANC/wither height].
A total of 30 mL of whole blood was collected from each horse, with 20 mL into lithium heparin coated tubes, and 10 mL into EDTA coated tubes.a Blood samples were centrifuged within 15 minutes and plasma frozen at −80°C until analysis. Plasma glucose, insulin, leptin, methionine, 3-nitrotyrosine, dopamine (DOPA), superoxide dismutase (SOD), and erythrocyte glutathione peroxidase activity (RBCGPx) were analyzed from heparin collected plasma using published methods.[3, 12] Plasma NEFA and triglycerides were analyzed from EDTA collected plasma using published methods. Estimates of insulin sensitivity were calculated: the reciprocal of the square root of insulin (RISQI; [1/insulin]2) and the modified insulin-to-glucose ratio (MIRG; [insulin2/glucose]).
Statistical analyses were performed using statistical software.b In addition to BCG, horses were categorized by age (4–8 years, 9–12 years, 13–16 years, 17–20 years), sex (mares, geldings), and exercise (none, any level). Stallions (n = 9) were excluded from analyses that included the effect of sex. Plasma levels of all variables, except RBCGPx, were log10 transformed to ensure normality of residuals. The effects of BCG, age, sex, exercise, and two-way interactions on glucose, insulin, NEFA, triglycerides, leptin, RISQI, and MIRG were assessed by mixed-models ANOVA. Correlations between insulin and NCHR, ANC, BW, and BMI were determined. The effect of breed on insulin, RISQI, and MIRG, and the effects of age, BCG, sex, and two-way interactions on RBCGPx, methionine, SOD, nitrotyrosine, and DOPA were determined using mixed models ANOVA. For all analyzes, Dunnett's tests were used to compare significant main effect means, whereas pairwise contrasts were used to compare means of significant interactions, and Tukey tests were used to compare means when a main effect of breed existed. Significance was set at P < .05.
Breed representation consisted of mixed breed (n = 71), Quarter Horse (n = 54), Warmblood (n = 45) Thoroughbred (n = 42), Tennessee Walking Horse (n = 20), Arabian (n = 16), Rocky Mountain Horse (n = 12), Paint (n = 11), Appaloosa (n = 8), American Saddlebred (n = 7), Racking Horse (n = 4), Fjord (n = 3), Andalusian, Irish Sport Horse, Morgan Horse, Mustang, Palomino, Paso Fino, and Spanish Mustang (n = 1 for each). There were 140 mares, 151 geldings, and 9 stallions. There were 106 horses (35.3%) aged 4–8 years, 76 (25.3%) aged 9–12 years, 68 (22.7%) aged 13–16 years, and 50 (16.7%) aged 17–20 years. Exercise regimens for 9 horses (3.0%) were categorized as intense, whereas 46 (15.3%) horses had moderate, 70 (23.3%) horses had light, and 175 (58.3%) horses had no exercise. Two hundred sixty-four horses had continuous access to forage at the time of sampling (253 horses had continuous access to pasture only, 7 horses had continuous access to pasture and hay, and 4 horses had continuous access to hay only), whereas 36 horses had restricted access to forage before sampling.
Geometric means of plasma variables and mean RBCGPx of the 300 horses are presented in Table 1. Two hundred forty-five (81.7%), 25 (8.3%), 9 (3%), and 21(7%) horses had insulin levels <20, 20–30, 30–40, and >40 mU/L, respectively (Table 2) of which 106 (43.2%), 19 (76%), 8 (88.9%), and 20 (95.2%), respectively, were overcondition or obese.
|Variable||GM||Lower 95% CI||Upper 95% CI|
|MIRG, (mUinsulin2)/(dL × mgglucose)||4.14||3.87||4.42|
|RBCGPx, U/g protein||242.3||231.1||253.6|
|SOD, U/g protein||100.4||90.0||112.1|
|Under||6 (2.5%)||0 (0.0%)||0 (0.0%)||0 (0.0%)|
|Optimal||133 (54.3%)||6 (24.0%)||1 (11.1%)||1 (4.8%)|
|Over||78 (31.8%)||9 (36.0%)||5 (55.6%)||5 (23.8%)|
|Obese||28 (11.4%)||10 (40.0%)||3 (33.3%)||15 (71.4%)|
|Total||245 (81.7%)||25 (8.3%)||9 (3%)||21 (7%)|
Two hundred and ninety-five horses were used for the statistical analyses that included the BCG variable, the 5 removed for having a BCS <4 included 3 mares (1 Saddlebred aged 17 years, 1 Quarter Horse aged 4 years, 1 Thoroughbred aged 19 years) and 2 stallions (1 mixed breed aged 8 years, 1 Thoroughbred aged 19 years). Plasma glucose levels were influenced by BCG (P < .001), whereby glucose (geometric mean, lower 95% confidence interval, upper 95% confidence interval) was higher in obese (97.5 mg/dL, 95.3, 99.9; P < .001) and overconditioned (94.1 mg/dL, 92.7, 95.5; P = .019) than optimal condition (91.7 mg/dL, 90.6, 92.9) horses. Age, sex, and exercise level had no effect on glucose levels (P > .2). Body condition group (P < .001) and age (P = .002) were associated with plasma insulin, RISQI and MIRG (Fig 1). Overconditioned and obese horses had greater insulin (P < .001) and MIRG (P < .001) and lower RISQI (P < .001) than optimal condition horses. Horses aged 17–20 had higher insulin (P = .002) and MIRG (P = .002) and lower RISQI (P = .002) than younger horses. Exercise level and sex were not associated with insulin, RISQI, or MIRG (P < .01). Levels of NEFA were lower in exercised (103.0 μEq/L, 82.8, 128.2) than nonexercised (118.6 μEq/L, 90.7, 155.1) horses (P = .002), and were not influenced by BCG, age, or sex (P > .4). Triglyceride levels were greater in mares (33.4 mg/dL, 30.1, 37.0) than geldings (27.9 mg/dL, 25.1, 31.0; P = .014), and in obese (36.3 mg/dL, 30.4, 43.3) comparedwith optimal condition (26.8 mg/dL, 24.2, 29.5) horses (P = .006). Triglyceride levels were similar in overcondition and optimal condition horses (P > .3), and were not associated with age or exercise (P > .1). A BCG by sex interaction (P = .047) existed for plasma leptin, whereby, within sex, overconditioned and obese horses had greater leptin levels than optimal condition horses (P < .01; Fig 2). Further, overconditioned mares had greater leptin levels than overconditioned geldings (P = .033). The BCG by age interaction was also significant (P = .033) for leptin, whereby mares aged 4 to 8 (P = .019) and 13 to 16 (P = .010) had greater leptin levels than geldings (Fig 2).
Plasma insulin was positively correlated with NCHR (r = 0.307, P < .001), ANC (r = 0.146, P = .013), and BMI (r = .316, P < .001), but not BW (r = .086, P > .1; Table 3). Breed affected plasma insulin (P = .006) and RISQI (P = .005),but not MIRG (P = .077), whereby insulin levels were greater and RISQI calculations were lower in Rocky Mountain Horses than mixed breeds, Quarter Horses, and Thoroughbreds (P < .05; Fig 3).
Mean RBCGPx activities were associated with age (P = .030) and BCG (P = .001), whereby obese horses had lower values (P = .002) than optimal condition horses, and horses aged 17 to 20 had lower values (P = .012) than horses aged 4 to 8 (Fig 4). There was no effect of sex on RBCGPx (P > .6). Methionine levels were not influenced by age, BCG, sex, or two-way interactions (P > .1). Eight horses (4 mares, 4 geldings) that ranged in age from 6 to 18 years, and in BCS from 6 to 9, had values higher than 1,136 and up to 50,978 nmol/mL, considered outliers for nitrotyrosine. The remaining values (n = 280) ranged from 50 to 464 nmol/mL, and were influenced by sex (P = .021), with geldings having higher values than mares (Fig 4). Nitrotyrosine values were not influenced by age, BCG, or two-way interactions (P > .1). Levels of SOD were not influenced by sex, age, BCG, or two-way interactions (P > .2). Nine horses (5 mares, 4 geldings) that ranged in age from 6 to 18 years and in BCS from 5 to 9 had values greater than 197 and up to 12,393 nmol/mL, considered outliers for DOPA. The remaining values (n = 279) ranged from 45 to 116 nmol/mL and were not influenced by sex, age, BCG, or two-way interactions (P > .1)
Several factors have been identified as potentially influencing laminitis risk in obese horses and ponies, including elevated plasma insulin, leptin, and triglyceride levels, and oxidative stress.[3, 11-13] We performed an observational, cross-sectional study in order to evaluate the relationship of these risk factors with obesity in a population of mature horses in Virginia. We report that overconditioned and obese horses had greater insulin and leptin concentrations than optimally conditioned horses, and that obese horses had greater triglyceride levels and lower red blood cell gluthathione peroxidase activities than optimal condition horses.
Obesity is associated with both HI and IR, with insulin levels increasing concurrently to weight gain. In this study, horses with a BCS of 7–9 had elevated insulin levels and lower estimated insulin sensitivity compared with optimal condition horses.[11, 13] Similar associations between insulin and body condition have been previously reported, in which horses with a BCS of 4 to 6 had lower insulin levels than horses with a BCS of 7 to 9. In addition to obesity, we found that horses aged 17–20 had higher insulin levels and lower insulin sensitivity estimates than younger horses. In a population of horses aged 3 to 29 years, age was negatively correlated with insulin sensitivity, and this relationship was not influenced by body condition, indicating that older horses might be at a higher risk of insulin resistance even if nonobese. The lack of a BCS by age interaction in this study supports this, although this interpretation might be limited by the number of horses in each group.
In addition to insulin, elevated leptin levels are a potential laminitis risk factor. Leptin, an adipocyte hormone, indicates energy balance to the brain, and plasma levels increase concurrent to weight gain, suggesting that leptin is responsive to short-term changes in body condition.[13, 15] Interestingly, plasma leptin is also increased by short-term infusion of insulin in humans and horses.[16, 17] Therefore, in this study, we are unable to attribute increased leptin levels in overcondition and obese horses solely to the horses' greater BCS, because they also had elevated insulin levels. Overconditioned mares had higher leptin levels than geldings, although in a previous report stallions and geldings had greater leptin levels than mares. In this study, horses were not balanced for BCS or age, and this could partially account for discrepancies with previous research.
In this study glucose levels were higher in the overconditioned and obese horses, a finding that is similar to those of a previous study reporting higher glucose levels, but within the reference range, in overconditioned and obese horses. In other species, elevated glucose levels in an insulin-resistant individual can indicate the onset of type II diabetes; however, this disease occurs infrequently in horses. In this study, glucose values ranged from 73 to 126 mg/dL, and although these values are slightly higher than those previously reported for obese nonfasted horses, they are within the range for healthy horses. As access to forage before sampling could have influenced the glucose values, the effect of body condition on glucose should be investigated in fasted horses.
Plasma triglyceride levels are also increased in obese horses, a finding that is possibly because of prolonged IR.[11, 13] Hepatic synthesis and release of TG as VLDL is normally decreased by insulin; however, the ability of insulin to turn off TG synthesis is reduced during insulin resistance, leading to elevated blood TG levels. In the present experiment, only obese horses exhibited elevated triglyceride levels, which could be a result of greater magnitude or longer duration IR. Triglyceride levels were also greater in mares than geldings, which could indicate that lipid metabolism is influenced by sex. We found that NEFA levels were similar across BCS groups, which opposes a previous report that plasma NEFA are elevated in overconditioned and obese horses. Having access to forage may preclude detection of BCS effects on NEFA levels. Interestingly, horses receiving any amount of exercise had lower NEFA levels than horses receiving no exercise. In humans, even moderate bouts of exercise increase insulin sensitivity, fat oxidation, and improved blood lipid profiles. In horses, exercise training improved insulin sensitivity when weight loss occurred concurrently, but without weight loss there was no improvement in insulin sensitivity.[22, 23] Duration of time spent at the indicated exercise level was not recorded for this study.
Although assessment of body condition is the most commonly used method to determine obesity, other morphometric measurements, such as neck circumference and body weight:height ratios, might be less subjective. In this study, plasma insulin levels were positively correlated with ANC, NCHR, and BMI measurements. Neck circumference measurements, ANC and NCHR, take into account subcutaneous fat deposits along the crest of the neck, an area of fat related to laminitis risk. Both ANC and NCHR positively correlate with insulin levels. The BMI system calculates a weight to height ratio, thus indicating obesity when horses at the same height have a higher BW. As expected, BMI but not BW alone correlated with plasma insulin in this study.
Anecdotal evidence and limited research reports suggest that certain breeds have greater predisposition to IR and laminitis. For instance, at an equine hospital specializing in first opinion and referral cases, presentation for laminitis was more prevalent in Icelandic horses than Warmblood and Light Breed horses, whereas Icelandic horses had greater insulin responses to hay feeding than Standardbred horses in a controlled feeding study.[4, 26] In this study, insulin levels and estimates of IR were compared between breeds represented by at least 10 individuals. Interestingly, Rocky Mountain Horses had greater insulin levels and lower RISQI calculations than Thoroughbreds, Quarter Horses, and mixed-breed horses. It is possible that BCS influenced these results, as 10 of the 12 Rocky Mountain Horses were overconditioned or obese, whereas only 5 of 42 Thoroughbreds, 32 of 54 Quarter Horses, and 40 of 71 mixed-breed horses were over or obese condition. Future studies should compare breeds balanced by BCS to further investigate the effect that breed may have on IR.
Increased oxidative stress and associated tissue injury, because of an imbalance in the production of reactive oxygen species or reduced capacity of antioxidants, has been proposed as a possible mechanism for laminitis. Reactive oxygen species have a role in regulating endothelial function and vascular constriction and relaxation, and thus could be involved in the pathogenesis of laminitis. In this study, RBCGPx activities were lower in both obese and older horses, potentially indicating that these groups have reduced capacity to metabolize reactive peroxides to inert forms. In ponies, RBCGPx activities were not associated with age, nor were they lower in ponies with a history of laminitis. However, it is possible that pony breeds are not an adequate comparison for horse breeds, and in this study no other oxidative stress markers were altered by obesity or age. Values reported for horses in this study were higher than values reported for ponies. Nitrotyrosine is a product of reactive nitrogen species-induced nitrative stress, and higher levels are often observed in diseased tissues and may have a pathologic role in the development of diabetes in humans. In this study, nitrotyrosine levels were greater in geldings than mares, indicating that geldings in this study may have had greater nitrative stress, but whether this relates to a disease process is unknown. Potential confounding factors for this study are that diets and access to forage before sampling varied. We attempted to minimize the effect of diet at the time of sampling by controlling feeding as best possible without discouraging owner participation. Caretakers did not feed horses concentrate after 2000 hours the night before sampling. Horses that were stalled overnight were not given more than 2 flakes of hay after 2000 hours the night before sampling and not turned out before sampling. Horses that were kept on pasture overnight or that had continuous access to hay (ie, round bale hay in a dry lot) or both pasture and hay (ie, round bale hay in the pasture) were allowed to continue to consume forage before sampling. We believe this approach allowed the horses to be maintained in as near their normal routine as possible while reducing the likelihood of consumption of feed just before sampling.
The present report expands our knowledge about the relationship between obesity and plasma insulin levels in equids. The novelty of the current data set lies in the large number of horses included (n = 300) whereas previous data sets have included fewer horses (n = 60; n = 46) or utilized ponies (n = 66) Our finding of higher insulin and leptin levels in overconditioned and obese horses agrees with previous evidence, and also provides support that these effects occur across sex, breed, and age. Of further significance is our finding of reduced RBCGPx activity in both obese and older horses, which could suggest a reduced capacity of these horses to modulate oxidative stress.
This study was supported in part by the Virginia Horse Industry Board, Virginia Tech [Virginia-Maryland Regional College of Veterinary Medicine and the College of Agriculture and Life Sciences], and the Macromolecular Interfaces with Life Sciences (MILES) Integrative Graduate Education and Research Traineeship (IGERT) of the National Science Foundation under Agreement No. DGE-0333378. The authors thank Kimberly A. Negrin, Julie Franklin, and Louisa Gay for data collection, Louisa Gay and James Martin for technical laboratory expertise, and Stephen R. Were for statistical consultation and analyzes.
Conflict of Interest Declaration: Dr. Geor currently is an Associate Editor with the Journal of Veterinary Internal Medicine.
BD Vacutainer®, Beckton, Dickson, and Company, Franklin Lakes, NJ
SAS Enterprise Guide v 4.2, SAS Institute, Carey, NC