Because diabetes involves imbalances of various substances in your blood like glucose and insulin, diabetes tests are often concerned with sampling your blood and analyzing its contents. Perhaps the most basic example is a blood glucose test, which simply reports the concentration of sugar in a patient’s blood at the moment of sampling. This information is useful but does not offer a complete picture of a diabetic’s condition.

This is where other tests like A1C come into play. A1C testing offers broader insights into trends in blood sugar levels that moment-to-moment blood glucose testing does not. However, as we will discuss, A1C testing can be of limited use in some situations. Other tests have been created to provide estimates of A1C test results, such as the GMI method, which may not suffer from the same limitations as A1C. This article will explore some recent insights into the accuracy of GMI methodology, and its viability as an A1C alternative.

What is A1C Testing?

A1C TestA1C, also known as Hemoglobin A1C or HgbA1c, is a molecule that helps red blood cells transport oxygen around the body. While this may seem irrelevant in the context of diabetes, A1C has some secondary properties that make it useful for assessing approximate long-term blood sugar levels. Specifically, A1C tends to become coated with sugar from the surrounding bloodstream, and consequently the concentration of red blood cells with sugar-coated A1C molecules increases or decreases in proportion to longer-term blood sugar concentrations.

This property allows practitioners to estimate a person’s average blood sugar levels over the course of several months, simply by drawing a single blood sample and determining the percentage of coated A1C molecules it contains. In fact, A1C concentrations can be directly converted to estimated average glucose (eAG) values, in the same units as blood glucose test results (mm/dL).

Some level of A1C coating is normal and expected in all humans since all humans should have some concentration of sugar in their blood. In fact, having up to roughly 5.7% of A1C coated with sugar is completely normal and suggests an absence of diabetes. Concentrations above 5.7% begin to suggest that an individual may be prediabetic, and concentrations above 6.5% almost certainly indicate full-blown diabetes, making this test an important part of a diabetes screening panel.

Regular A1C testing can also help diabetic patients track their conditions over time. This is because A1C levels in diabetics have been correlated to the incidence of future diabetes complications. Diabetic individuals and their doctors may therefore set A1C targets or goals depending on their circumstances. A goal of 7% is typical, though individuals should always follow the advice of their physician.

It’s important to stress that A1C tests are merely estimates of average blood glucose, which will inevitably carry some degree of inaccuracy. Such inherent inaccuracy is largely due to the fact that several factors unrelated to diabetes can impact hemoglobin levels, such as kidney failure, ethnic background, and anemia. This inherent inaccuracy is one reason that A1C tests should only form part of a larger health assessment plan for diabetics, and should complement, not replace other tests.

What is GMI and How Does it Relate to A1C?

Just as there are many paths to the top of the proverbial mountain, so too are there many ways to assess long-term trends in blood glucose levels. Like A1C testing, GMI is a methodology that aims to provide information about a person’s historical blood sugar levels, albeit through alternative methods.

GMI, which stands for Glucose Management Indicator, is simply an estimate of a person’s A1C levels. Instead of sampling A1C concentrations to find this value, GMI is derived from a person’s mean glucose levels over 14 days or more and converted to an A1C estimate based on empirical data from large population samples. Mean glucose levels can be obtained through regular blood testing, or through a Continuous Glucose Monitor, which regularly and automatically assesses a person’s blood glucose levels.

Advantages of Using GMI to estimate A1C

A1C and GMI testing are similar in many respects. Both tests require blood samples, both assess the concentration of biological markers in the blood, and both provide an estimate of long-term trends in blood sugar levels. However, while the tests serve similar purposes, they may not be equal in their accuracy or reliability.

As we discussed previously, A1C testing can fall short due to various factors unrelated to blood sugar or diabetes. This can make it a somewhat unreliable indicator of long-term blood sugar trends since it’s not always possible for doctors to control for the presence of such complicating factors as anemia or kidney disease. GMI circumvents this issue by testing blood sugar directly and averaging the results over time to cancel the effects of atypical readings or outliers. This can make GMI a more effective assessment tool for people to whom the previously discussed factors apply. Intuitively, this makes sense: after all, A1C is meant to represent average blood sugar, so why wouldn’t actual measurements of average blood sugar levels be more accurate?

Perhaps one drawback of GMI is that it requires constant sampling over a long period of time to obtain the necessary data, rather than a single blood sample (as is the case with A1C testing). This is not likely to be a significant issue for many diabetics since diabetic individuals are often in the habit of regularly testing their blood glucose or employ the use of a continuous glucose monitor (CGM).

Recent Insights into Accuracy of A1C Estimates

With the advantages of GMI looking pretty clear, perhaps the only remaining question is how accurately CMI can estimate A1C. Until recently, little data existed to address this question. However, a recent study  from the University of Washington reviewed historical data to determine how closely the GMI-estimated A1C values and actual measured A1C values for hundreds of diabetic patients have been correlated.

The results of their study were illuminating. They found that a whopping 22% of the study population reported differences between estimated and actual A1C of 1% or greater. This may seem insignificant but recall that the difference between a healthy A1C percentage and a diabetic A1C percentage can be as little as 0.8%.

Slightly more reassuring were their findings that 50% reported differences of 0.5% or less, with 11% showing less than 0.1% difference. Also, unsurprisingly, the authors reported that patients with kidney disease were the most likely to report more significant differences. This may suggest that GMI is still a relatively reliable A1C estimator for individuals without any of the relevant applicable factors.

The authors also noted that their study had some inherent limitations. Because they were analyzing historical data, instead of setting up and conducting an observational study, the authors were unable to control for a number of outside factors like changes in medication or recent diagnoses of illnesses. The population of patients analyzed was also relatively ethnically homogenous, which might limit the applicability of the results (given that A1C values can be artificially high or low for people from some ethnic backgrounds).

Overall, this study provides valuable insight into the accuracy of GMI, but should not be seen as a conclusive endorsement for or against GMI. For now, GMI will likely remain another tool in the arsenal of doctors and their patients to keep tabs on the progress of diabetes and blood sugar levels over time.

Alternative Tools for Longer-Term Tracking

When comparing the virtues of GMI and A1C, it can be easy to forget that there are other ways to track the effectiveness of treatment. In fact, GMI and A1C both share shortcomings that other methods can overcome.

For example, a problem that is common to GMI and A1C testing is the inability to capture short-term fluctuations that might fall well outside of a healthy blood sugar range. A person could, for example, spend half their time with dangerously high levels of blood sugar, and the other half with dangerously low levels of blood sugar. The average could still balance out somewhere near their target GMI or A1C level.

To counteract this problem, metrics like Time in Range provide a breakdown of the number of CGM readings that fell inside, above, or below the target blood sugar range for a person over a period of time. This information can serve a similar purpose as knowing the average blood sugar levels but has the added bonus of reporting the time a person spent with unhealthy blood sugar levels.


If nothing else, this discussion should illustrate the wide variety of analytical tools available to track how well your blood sugar is being controlled. Each technique has its own strengths and weaknesses, and together they can provide a broader and more nuanced picture of the progress of your treatment. GMI and A1C are two such options that are intended to boil down your blood glucose levels into a single number. Each has limitations, and neither should be taken as a conclusive indicator of your health or the efficacy of your treatment.

Remember, diabetes treatment is a tricky business, and there’s a lot to keep track of. While it’s good to understand the terminology used by your healthcare professionals to facilitate better communication between you and them, it’s important to rely on their guidance regarding your treatment. They will determine which tests you should be subject to, and how often. If you have concerns about how your diabetes is being treated or assessed, be sure to raise it with them at the next availability.