What is insulin resistance?
Insulin resistance is classically defined as impaired whole-body insulin-mediated glucose disposal.1
However, defective insulin action is not confined to impaired glucose disposal; it can also be seen in protein and fat metabolism. Therefore, a more accurate definition of insulin resistance is “a state in which normal insulin levels produce an attenuated biological response.”1
Improvement of insulin sensitivity/insulin action, or reduction in insulin resistance, as a therapeutic principle may play a central role in the primary prevention and treatment of type 2 diabetes, hypertension, and atherosclerosis.2
How common is insulin resistance?
Insulin resistance is very common in the general population. The prevalence of insulin resistance has been estimated to be as high as 88% in people with lipid abnormalities, such as elevated triglycerides and low HDL.3 Considering that over half of all adults in Western populations have lipid abnormalities,4 it becomes clear that insulin resistance is not a problem confined to diabetes but a mass problem afflicting the vast majority. A surprising finding is that people with elevated triglycerides and low HDL can have a similar degree of insulin resistance as that seen in patients with type 2 diabetes.3
Importantly, in the general population insulin resistance can be found even in the absence of any major metabolic disorders,3 and 25% of apparently healthy people are severely insulin resistant.5 This can be detected by elevated fasting insulin and glucose levels 6,7 – and the HOMA-IR index (explained below) - and may be an indicator of “silent” atherosclerosis.8
What is the value of assessing insulin resistance?
Insulin resistance is an independent risk factor for all-cause and cardiovascular mortality,9,10 and multiple studies have shown that insulin resistance is a strong predictor of atherosclerotic cardiovascular disease as well as adverse cardiovascular outcomes in both non-diabetic and diabetic individuals.10,11
A notable study found that insulin resistance, as estimated by HOMA-IR (explained below), was associated with incident symptomatic cardiovascular disease after a 15-year follow-up.12 Importantly, the association of insulin resistance with cardiovascular disease was independent of classic risk factors (including hyperglycemia, hypertension, high LDL cholesterol, smoking, and physical activity) and of other components of the metabolic syndrome (obesity, hypertriglyceridemia, and low HDL cholesterol). The association remained significant and virtually unchanged even after accounting for novel risk factors, including adiponectin and biomarkers indicating a prothrombotic state (high fibrinogen), increased oxidative stress (high circulating oxidized LDL), endothelial dysfunction (high VCAM-1), and chronic mild inflammation (increased hsCRP).12
Furthermore, insulin resistance causes physiological impairments such as telomere attrition and endothelial dysfunction, which contribute to premature aging.13-16 Experimental studies have shown that insulin resistance accelerates atherosclerosis.17,18 In apparently healthy people, 1 out of 3 who have insulin resistance at baseline develop either heart disease or type 2 diabetes 6 years later, while people without insulin resistance do not.19 Insulin resistance seems to be the common metabolic defect underlying both type 2 diabetes, dyslipidemia, cardiovascular disease (including high blood pressure) and obesity.20,21 With the rapidly growing body of evidence linking insulin resistance to adverse health outcomes and premature death, the interest in practical ways to assess insulin resistance is escalating.
HOMA-IR, an index of insulin resistance
Insulin sensitivity / insulin resistance can be assessed in several ways. The hyperinsulinemic euglycemic glucose clamp technique is the gold standard method for the assessment of insulin sensitivity, because it directly measures the ability of insulin to promote glucose uptake into tissues from the blood.22
During the clamp procedure, intravenous infusions of insulin and glucose are given. The principle of the test is to keep the glucose level constant while increasing insulin levels.22 This is a very resource intensive method that is only conducted in research settings.
For large scale studies and in clinical practice, several simple surrogate assessments of insulin sensitivity / insulin resistance are available, the most popular of which is the homeostatic model assessment (HOMA) of insulin resistance (IR).23,24 HOMA-IR is a simple calculation based on measurements of insulin and glucose in a fasting blood sample. The formula for HOMA-IR is:
HOMA-IR = (FPI x FPG)/22.5
Where FPI is fasting plasma insulin (mU/L or µU/mL) and FPG is fasting plasma glucose (mmol/L).
The denominator of 22.5 is a normalising factor, i.e., the product of normal fasting plasma insulin of 5 µU/mL and normal fasting plasma glucose of 4.5 mmol/L, obtained from an apparently healthy “normal” individual.23 Therefore, for an individual with normal insulin sensitivity, HOMA = 1. Surveys of representative populations confirm that lean and apparently healthy individuals have HOMA-IR values between 0.5 and 1.25,26
Why is insulin alone not a good indicator of insulin resistance?
Up to 75% of people with normal fasting glucose / glucose tolerance may have hyperinsulinemia, which is the hallmark adaptation to insulin resistance.27 In healthy populations with normal glucose levels, fasting insulin alone may be a good predictor of insulin resistance, atherosclerosis and cardiometabolic risk.28,29 However, in populations with overweight/obesity, metabolic syndrome, type 2 diabetes, cardiovascular disease and non-alcoholic fatty liver disease, HOMA-IR is a better indicator of insulin resistance and metabolic disease outcomes than insulin levels alone.30,31
HOMA-IR, by including both glucose and insulin, is more generalizable to all circumstances with variable glucose levels, including prediabetic glucose levels commonly seen in people with obesity and/or visceral fat accumulation.30,32 By accounting for a wide range of metabolic conditions, HOMA-IR is a better indicator of insulin resistance than insulin levels alone, and has a greater correlation with the gold standard clamp method among obese and type 2 diabetic subjects.30,32 The superiority of HOMA-IR over insulin alone was demonstrated in a meta-analysis which showed that insulin resistance assessed by HOMA-IR, but not fasting insulin, is independently associated with greater risk of cardiovascular and all-cause mortality.10
How to interpret HOMA-IR values
The higher the HOMA-IR value, the greater the degree of insulin resistance. Hence, individuals with type 2 diabetes have much higher HOMA-IR values (often in the range of 5-6 and above) than weight-matched individuals without type 2 diabetes (often below 2).33
Numerous studies have investigated the association between insulin resistance defined by HOMA-IR and various health outcomes, as well as identified thresholds to identify subjects at increased risk for the metabolic syndrome, prediabetes and type 2 diabetes. However, it has to be kept in mind that these HOMA-IR thresholds only apply to the population from which they were derived, and hence the thresholds that identify people at high risk vary between populations.34 It is also important to point out that direct comparisons of HOMA-IR values between studies cannot be made unless the same insulin assay was used. Nevertheless, below is a summary of notable studies that investigated the association of HOMA-IR with various outcomes. These studies provide a rough guideline about the range of HOMA-IR values that can be expected to be seen in patients encountered in clinical practice.
One study found that the optimal HOMA-IR cut-off for the identification of the metabolic syndrome (IDF- and ATPIII-defined) in individuals aged 25-64 years was 1.77.35 Similarly, another study found that HOMA-IR of 1.85 identifies men with cardiometabolic risk factors.36 Optimal cut-off points to identify men with incident type 2 diabetes was HOMA1-IR = 2.17 and HOMA-B% = 67%.37 HOMA-IR values around 2-2.5 show diagnostic value in distinguishing non-alcoholic fatty liver disease carriers from control group individuals.38
Importantly, there is a significant association between HOMA-IR and risk of cardiovascular disease even after adjustment for multiple confounders.39 For example, in the San Antonio Heart Study, the risk of a cardiovascular event increased by 2.5-fold across increasing HOMA-IR categories, even after adjustment for age, sex, and ethnicity. After additional adjustment for LDL, triglyceride, HDL, systolic blood pressure, smoking, alcohol consumption, exercise, and waist circumference, individuals with HOMA-IR = 7.3 had a 2-fold higher risk of a future cardiovascular event compared to individuals with HOMA-IR = 0.6.39
One meta-analysis found that people with the highest vs. lowest HOMA-IR category had 34% higher risk of all-cause mortality and 2.1-fold higher risk of cardiovascular mortality.10 The HOMA-IR categories that were compared in included studies in the meta-analysis were 2.5 vs. 0.6,40 >2.8 vs. ≤1.4,41 >1.5 vs. 0.85–1.07,42 and >2.67 vs. <1.29.43
The research on HOMA-IR demonstrates that the diagnostic cutoffs for conditions related to insulin resistance vary between populations and studies. Nevertheless, when used and interpreted correctly, HOMA-IR can add value in clinical practice, as will be explained next.
How to use HOMA-IR – the value of assessing HOMA-IR in routine clinical practice
Assessment of insulin resistance with the HOMA-IR tool can be used for three purposes:
- To evaluate baseline degree of insulin resistance.
- To monitor progression of insulin resistance over time.
- To monitor the efficacy of insulin sensitizing treatments.
Due to the variability in insulin assays,44,45 and pulsatile nature of insulin secretion (which is maintained after an overnight fast, although less insulin is secreted with every burst) 46, there are no universal diagnostic thresholds for “normal” vs. “abnormal” HOMA-IR values.26 Nevertheless, as outlined below, numerous studies have compared metabolic status in subjects with high vs. low HOMA-IR values, and quantified the association of high vs. low HOMA-IR values with risk for various outcomes.
Rather than making the “diagnosis of insulin resistance”, the greatest utility of HOMA-IR in clinical practice is for monitoring progression / regression of insulin resistance in patients over time, and evaluating the response to insulin sensitizing treatments. When using HOMA-IR for this purpose, it is critical that the same insulin assay is used, and that patients do the fasting blood draw under the same conditions each time (e.g. drawing blood around the same time, overnight fasting for the same number of hours and eaten the same meals the day before the blood draw, as they did before the previous blood draw).
In a given population without type 2 diabetes, the interindividual and intraindividual variation of HOMA-IR values is relatively low.47 In contrast, in patients with type 2 diabetes there is a larger variation in HOMA-IR values. The high degree of intraindividual variability of HOMA-IR seen in individuals with type 2 diabetes means that HOMA-IR values must increase by 90% or decrease by 47% in order to conclude that a significant change in insulin resistance has occurred; smaller changes in HOMA-IR values may simply reflect intraindividual variability.47 Nevertheless, HOMA-IR was found to be a reliable indicator of insulin resistance during follow-up of patients with type 2 diabetes,48 and can be used to examine changes in insulin resistance in response to various treatments,49,50 including diet/exercise 50,51 and testosterone therapy.52-57
For instance, a 1-year long lifestyle intervention reduced HOMA-IR from 1.5 to 1.0 in subjects with a modest degree of baseline insulin resistance, and from 4.6 to 2.7 in subjects with a greater degree of baseline insulin resistance.51 Men with hypogonadism commonly have HOMA-IR values in the range of 4-16, indicating a relatively severe degree of insulin resistance, which is effectively reduced by testosterone therapy.52-57 For example, one notable randomized, double-blind, placebo-controlled trial (RCT) found that testosterone therapy for 2 years in men with obesity reduced HOMA-IR from 4.27 to 2.17.52 Interestingly, this was accompanied by a significant 23% reduction in carotid intima-media thickness (CIMT),52 a surrogate measure of atherosclerosis.58 Another RCT in men with obesity and type 2 diabetes showed that testosterone therapy for 1 year reduced HOMA-IR from 11.45 to 6.81, along with a 11% reduction in CIMT.56 These studies suggest that testosterone therapy reduces both insulin resistance and the atherosclerotic burden in men with hypogonadism, regardless of type 2 diabetes status.
Implementing regular assessment of insulin resistance in clinical practice is warranted for numerous reasons, including:
It is useful for the clinician to have an objective measure of the improvement in insulin resistance in response to diet/exercise and insulin sensitizing treatments, such as testosterone therapy. The presence of a high degree of insulin resistance - as indicated by elevated HOMA-IR values - could prompt an upward dose titration and more frequent monitoring of testosterone levels to ensure that optimal testosterone levels are achieved during testosterone therapy, and/or escalation of other insulin sensitizing treatment(s). This in turn could help reduce residual cardiovascular risk.