Diagnosing Insulin Resistance: Q&A with Researchers
Over the past few years researchers have described a strong association between insulin resistance and laminitis in equines. They are working now on defining standard testing protocols and interpretations to identify horses at highest risk for laminitis. Many questions remain unanswered. How should insulin resistance be defined and diagnosed? How do researchers interpret test results? Can blood tests alone determine the risk of our horse or pony to get laminitis? Until they have more solid science to configure a standard definition of equine insulin resistance, those attempting to define it might find themselves in the same predicament as the proverbial group of blind men describing an elephant.
A study in the United States showed that laminitis affected 2% of all horses, with the incidence going up to 5% in spring, which is when grass sugars peak. The ability to identify high risk animals before laminitis strikes is essential, as this can allow caretakers to implement appropriate management practices to prevent it. Sinking or rotation of the coffin bone requires treatment and rehabilitation regimes that can be difficult, long, expensive, and emotionally draining. Even then, the treatments often fail. But how do we identify the high-risk animal? Of course any horse that has had a brush with laminitis should have tests for underlying endocrine problems, but certain physical characteristics should lead proactive owners to test for insulin resistance.
A "cresty neck" is the classic sign of insulin resistance, and researchers note a solid correlation between neck circumference and the condition. Horses that gain weight much more quickly, often described as "blowing up on grass" than their herdmates under similar management, might be candidates for testing, especially if they also exhibit signs of foot tenderness. Additional signs are rings on the hoof wall, or a stretched white line with blood specks when the foot is viewed from the bottom, both which could be indications of a previous mild case of laminitis.
Not all fat horses are insulin resistant (IR), and not all IR horses are fat. Typical body scoring parameters might not apply to some IR horses. Even horses with ribs showing might have insulin resistance. Even when they maintain proper body weight with dietary restriction, IR equines might still display subtle signs of regional adiposity, with fat pads that form behind the shoulder, around the tailhead, over the loin, or as a slight crest with a dip in front of the withers. These abnormal fat deposits might be lumpy or dimpled in appearance.
In a clinical setting, dynamic tests are considered the most accurate for diagnosis of impaired endocrine function. The combined glucose-insulin tolerance (CGIT) test measures the horse's response to a standard dose of glucose and insulin over several hours with multiple blood draws. Insulin resistant horses maintain high glucose and insulin levels longer than normal. Compared to tests based on a single blood sample, the CGIT is not as practical in the field, is more expensive, and if the horse is stressed by being in a stall away from home, it might affect the results. Therefore, it becomes important to derive as much information as possible from diagnostic tests that can be done with a single blood draw. Baseline insulin and glucose levels can be very informative.
Although the scientific community is far from having all the answers, two leading researchers have shared their expertise and opinions about how insulin resistance should be diagnosed.
Nicholas Frank, DVM, PhD, Dipl. ACVIM, an associate professor of large animal internal medicine at the University of Tennessee is conducting studies on a herd of insulin resistant horses with chronic laminitis. Horse owners managing laminitic horses can relate to his description of this herd as "trying to keep a flotilla of leaky ships afloat."
Ray Geor, BVSc, MVSc, PhD, Dipl. ACVIM, director of research at Virginia Tech's Middleburg Agricultural Research and Extension Center was part of the team who first described pre-laminitic metabolic syndrome in a herd of Welsh and Dartmoor ponies. Ponies at greatest risk for laminitis could be identified by blood tests for insulin, glucose, and triglycerides even before their first laminitis episode was triggered by spring pasture.
Question: What is your stand on fasting before blood draw for insulin and glucose? Some say that the stress of fasting could alter test results, especially if the horse is used to having free access to hay.
Dr. Frank: There are two approaches that I recommend with respect to feeding and fasting before blood draws. If the horse remains calm when deprived of feed for a few hours, I recommend feeding a flake of hay at 10 p.m. the night before and then collecting blood the next morning. However, we have also have found that some of the obese horses that we test become very agitated when feed is withheld. This causes stress, which affects test results because stress hormones such as cortisol and epinephrine cause temporary insulin resistance. Stress during testing can, therefore, lead to the misdiagnosis of insulin resistance. We have not detected any short-term effects of feeding hay on resting blood glucose and insulin concentrations, so we recommend that the horse be fed hay as normal before and during testing. It is possible that feeding hay that is rich in sugars within the 12-hour period before testing could raise concentrations above fasting levels, but we choose to accept this possibility over the known complication of stress. Grain should never be fed in the 12 hours before testing.
Dr. Geor: Standardization of the sampling protocol is extremely important, particularly how the horse is fed during the 12-hour period prior to blood draw. Our studies have shown that the sugars in pasture forage can markedly impact blood insulin (and, to a lesser extent, glucose) levels, potentially leading to a false diagnosis of insulin resistance. This does not seem to be a problem when hay is fed, even when the nonstructural carbohydrate content is moderately high. Therefore, we recommend that horses are removed from pasture--to a stall or drylot-for about 12 hours prior to blood draw. No grain should be fed during this period but hay should be offered. Fasting is not a normal state for the horse and may cause stress that affects test results.
Are there times when testing for insulin and glucose is not appropriate?
Dr. Frank: The most important consideration before collecting blood is whether the horse is experiencing pain as a result of laminitis. If this is the case, stress hormones are likely to be causing insulin resistance and serum insulin levels will be high. It is, therefore, very important to wait until the pain of laminitis has subsided before testing horses for IR.
If a horse is tested on a hay-only diet, and blood levels of insulin or glucose are above the "normal" range of the testing laboratory, a positive diagnosis of hyperinsulinemia or hyperglycemia can be made. But what does "normal" mean? Different labs have different ranges for normal. The analytical procedure for insulin in horses is not standardized and, therefore, may differ from one lab to another. It is, therefore, important to select a laboratory that often measures insulin levels in horse blood and remain with the same lab to assess progress in an individual case.
Might a horse still be considered IR if test results for insulin are within the lab's "normal" range?
Dr. Frank: For our laboratory we consider a blood insulin level > 20 ï¿½U/mL (microunits per milliliter) to be suggestive of IR and > 30 ï¿½U/mL defines hyperinsulinemia and therefore insulin resistance. We occasionally encounter horses with normal insulin levels and abnormal intravenous combined glucose-insulin test (CGIT) results. These patients suffer from mild insulin resistance that is likely to be exacerbated by pasture grazing, grain feeding, or seasonal changes.
Dr. Geor: I do not place much faith in the reference ranges developed by some of the commercial labs. These ranges are sometimes based on very few animals. In general, I go with >20 ï¿½U/mL as suspicious and >30 ï¿½U/mL as more convincing evidence of hyperinsulinemia and an IR problem providing the sample was collected in an appropriate manner.
Please define "compensated" and "uncompensated" insulin resistance.
Dr. Geor: Compensated insulin resistance implies that glucose values are being regulated within the norm. There is insulin resistance but increased insulin secretion and, therefore, higher circulating insulin "overcomes" the insulin resistance, such that glucose is taken up by the cells and blood glucose is within reference limits. Uncompensated insulin resistance is when high insulin fails to overcome the resistance and blood glucose values rise above normal.
Do glucose levels alone give us useful information? What level defines hyperglycemia?
Dr. Frank: Most horses with insulin resistance are able to maintain blood glucose concentrations within reference range, so normal blood sugar levels are not diagnostic. Levels are rarely higher than the reference range (hyperglycemia), and when this problem is detected, it is cause for greater concern. Hyperglycemia often indicates that insulin resistance has been going on for long time and is seen in older patients. When hyperglycemia is detected, the veterinarian should test the patient for pituitary Cushing's disease because our research group has detected hyperglycemia in horses and ponies with this hormonal problem. I use 100 mg/dL, (milligrams per deciliter) but this is a somewhat arbitrary cut-off.
Dr. Geor: In the context of evaluating insulin resistance, it is necessary to evaluate both glucose and insulin, e.g., to differentiate compensated versus uncompensated insulin resistance. Blood glucose concentrations tend to rise very quickly in response to pain (e.g. due to current laminitis) or stress, so these factors need to be taken into consideration when interpreting blood glucose values. In general, a glucose value <100 mg/dl can be regarded as okay. Cushing's disease can result in increased glucose concentrations, so a finding of hyperglycemia (>100 mg/dl) in a horse or pony without evidence of current laminitis should prompt the veterinarian to test for this problem.
Might glucose: insulin ratios give us more information?
Although blood glucose levels usually remain within reference range, they may be higher than they would be if the horse were healthy. This is one reason for calculating the glucose to insulin (G:I) ratio. Unfortunately there are several problems with measuring blood glucose levels in horses. First, blood glucose levels quickly increase in the stressed horse, even within a few minutes. On the other hand, blood glucose levels decrease if the blood sample is handled improperly. Delaying the time between collection and centrifugation causes levels to decrease, particularly if the sample cannot be refrigerated. Finally, the major problem with the G:I ratio is that it does not account for differences between laboratories with respect to the insulin assay used. Glucose assays are very consistent between laboratories, but insulin values vary because different laboratories use different assays. If only insulin values are used to diagnose insulin resistance, results can be interpreted using the reference range provided by the laboratory. This reference range will have been adjusted up or down according to the assay used. However, you cannot account for differences in reference ranges when the G:I ratio is used, so horses look better or worse, depending upon the laboratory used.
Dr. Geor: For the reasons elaborated on by Dr. Frank, we do not use the glucose:insulin ratio for diagnosis of insulin resistance. Related to this question, there has been interest in the use of other proxy markers of insulin resistance based on measurements of glucose and insulin. The group at Virginia Tech developed proxy markers of insulin resistance (RISQI, calculated from blood insulin) and insulin secretory response (MIRG, modified insulin to glucose ratio) and used these to evaluate predisposition to laminitis in a herd of Welsh and Dartmoor ponies. Ponies with a history of laminitis were more insulin resistant as indicated by lower RISQI and higher MIRG compared with the ponies that were never laminitic. However, it must be stressed that the cutoff values used to separate the two groups should not be used for clinical diagnosis of insulin resistance in other populations of horses or ponies. For example, the RISQI cutoff represented an insulin value of 11-12 ï¿½U/mL--a value well below our clinically used cutoff for diagnosis of hyperinsulinemia.
Might a horse be considered IR based on a low G: I ratio even if the individual numbers are within normal range?
Dr. Geor: No, this would seem to be stretch and, in my view, will result in gross over diagnosis of insulin resistance.
Dr. Frank: The G:I ratio is used by some veterinarians to diagnose and characterize insulin resistance in horses. This value can be calculated by dividing the glucose concentration in mg/dL by the insulin concentration in ï¿½U/mL (same as mU/L). Glucose concentrations measured in mmol/L (millimoles per liter) must be multiplied by 18 to convert to mg/dL (milligrams per deciliter) and insulin concentrations reported in pmol/L must be divided by 7 to convert to ï¿½U/mL. For example, a horse with a blood glucose concentration of 80 mg/dL and serum insulin concentration of 5 ï¿½U/mL has a G:I ratio of 16.
Two cut-off values have been proposed for the interpretation of G:I ratios. The first is a value of 10, so a G:I ratio < 10 is supportive of IR. However, if we look at an example and set the blood glucose concentration at 90 mg/dL for the purposes of the discussion, a G:I ratio of 10 would be equivalent to a blood insulin level of 9 ï¿½U/mL, which is well within our reference range. A G:I value < 4.5 has also been proposed to diagnose insulin resistance, and this provides results that are more consistent with ours. If we go back to the example, the horse would have blood glucose and insulin concentrations of 90 mg/dL and 20 ï¿½U/mL, respectively. We suspect insulin resistance if the blood insulin concentration is > 20 ï¿½U/mL. Our research group does not use the G:I ratio because we have found that more variability is introduced when a value is based upon two variables, and each of these variables is potentially susceptible to sampling and handling error. Our group bases the diagnosis of insulin resistance on resting blood insulin concentrations, detection of hyperglycemia, or intravenous glucose and insulin challenge test results. I would not diagnose insulin resistance on the basis of a G:I ratio alone if blood insulin concentrations were not elevated.
Laboratory databases lump horse and ponies together. Is normal insulin for a pony the same as normal for a horse?
Dr. Frank: We currently do not distinguish between horses and ponies with respect to any of these parameters, although this work is in progress.
Dr. Geor: There is evidence that ponies are more insulin resistant than horses, potentially explaining their higher susceptibility to laminitis. Resting insulin values also tend to be higher in ponies compared to horses but we need more data before it will be possible to derive different cutoffs for diagnosis of insulin resistance.
Scientists hope research under way will someday help horse owners who suspect insulin resistance to get more accurate, standardized tests that can identify and define which horses require special management to avoid laminitis. The blood levels of insulin that consistently trigger changes in the foot are unknown. The mechanism for insulin induced laminitis is unknown. Scientists need to standardize the laboratory methods to quantify equine insulin, and they need to establish databases for "normal" insulin levels in populations with different genetic backgrounds.
Even if a horse is diagnosed unequivocally as being IR, that does not mean it will get laminitis. While insulin resistance as a background condition predisposes an animal to laminitis, avoiding the dietary and management practices that trigger the condition can prevent it from ever happening. This is why horse owners need to know.
Eliminating grain, increasing exercise, limiting access to high sugar pasture, and finding hays that test lower in sugar and starch can prevent laminitis even in animals that are predisposed.
Even if your horse or pony is borderline IR based on current definitions subject to revision, these practices generally can't hurt and might help a lot for a healthier future with a lot less worry.
Kathryn Watts, BS, is research director at Rocky Mountain Research & Consulting Inc.
POLL: University Equine Hospitals