Measuring Insulin Resistance

After doing so much research, lecturing and writing about insulin resistance, I have constant requests for more information on how to measure it in clinical practice. This is not an academic exercise: it is estimated that a person on the road to developing type 2 diabetes may have been insulin resistant for as long as twelve years before the disease is diagnosed.

In high-risk populations, there is a lot of value in regularly checking plasma glucose, but the problem is that once glucose begins to rise, it implies that the pancreas can no longer keep up with the demand for insulin and that we may be passing the point of no return.

These are the most common questions that I get::

1. Should you be having your insulin level measured?
2. Should you have your insulin resistance measured?
3. What’s normal?

First, measuring insulin levels themselves is not of much value: they bounce around a good deal in the course of a day, and many things can alter your circulating insulin levels.

Second, accurate measurement of insulin resistance is an expensive and cumbersome procedure involving intravenous sampling of blood and in some cases also giving intravenous insulin.

Third, there is no such thing as “normal.” Results derived from any kind of test are a “reference range.” This means that they show how a result related to a large group of apparently healthy people. This is an important concept. I often have students say, “What’s normal?” There is no such thing. Blood tests help and guide us but can only be understood in the context of the whole person.

We never treat a laboratory value: we treat people. You may be interested to have a look at an earlier article about this important issue.

But all is not lost: we do have a blood test that can be used to guide us. We don’t have evidence to suggest that we should be using it to screen the whole population for insulin resistance. Instead it is a test to help guide us in high-risk populations. The test is called the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR).

The original paper was published by a group of experts form the university of Oxford in 1985. The drawback of the HOMA-IR is that it is a mathematical model, and it’s only as good as the accuracy of an individual laboratory’s insulin assay.

Since then, the HOMA-IR has been used in epidemiological studies such as the famous Framingham study and there has been a lot of work on trying to correlate the HOMA-IR with other measures of insulin resistance. There are now over one thousand papers that reference it, and we have had a great deal of experience in using it in our studies of insulin resistance in people with mental illness.

Apart from research, we only use the HOMA-IR as a guide in high-risk individuals. A simultaneous fasting glucose and insulin are taken.

Insulin resistance (HOMA IR) =
Fasting insulin (µU/ml) X Fasting glucose (mmol/l) divided by 22.5.

Most studies now suggest that the cut-off for insulin resistance should be 1.7; although some have been slightly more forgiving, and suggested that up to 2.5 may be acceptable. But remember that the HOMA-IR is only giving us an estimate to help with the overall evaluation of a high-risk individual, and we do not treat a laboratory value.

If the value is above 2.5 many experts would suggest intervention if there are also features of the insulin resistance syndrome. The key interventions are diet and exercise, both of which have been proven to reduce insulin resistance. A very interesting approach adopted in two European studies has been to treat high risk people with a medication called metformin, and were able to show that within a year several cardiovascular risk factors improved.