top of page

Adverse Metabolic Effects of High Glycemic Meals: Part I- Glucose Variability

Two examples of daily blood glucose measurements. These are examples that are loosely based on our CGM data after eating well and cheat days.The green plot is indicative of a low carbohydrate, high fat diet and the red plot is an example of a high carbohydrate, high glycemic load diet (aka, the Standard American Diet (SAD)). Both green and red plots have the same average glucose of 97.6 mg/dL (est. HbA1c 5%). However, the green plot demonstrates low glucose variability with a standard deviation of 17 mg/dL while the red plot demonstrates high blood glucose variability with a standard deviation of 44 mg/dL. The following write-up is part I of a two part series examining the metabolic effects of consuming high glycemic load meals.


Laura and I have used our continuous glucose monitors (CGMs) to make lifestyle modifications with great success. We have identified a few supposedly healthy components of our diets that caused large reactive glucose spikes. Additionally, it has allowed us to design a tasty diet which causes minimal glucose variation. In this write-up I will discuss why minimizing glucose spikes are relevant to our healthy lifestyle, longevity, and healthspan by describing the deleterious physiologic response of the body to a large carbohydrate meal. The three main mechanisms are: induced postprandial hyperglycemia with resultant glucose variability, postprandial hyperinsulinemia, and postprandial reactive hypoglycemia. These effects will be divided into two write-ups.

Part I: Postprandial Hyperglycemia and Glucose Variability

The One Rep

  • The main mechanism of damage in high blood sugar is reactive oxygen species generation

  • Most of the risk with high blood sugar is due to sustained average blood sugar (HbA1c); however, there most likely exists some component that is due to glucose variability

  • Glucose variability is the amount by which glucose varies from the mean average glucose

  • The in vitro data on glucose variability is compelling; glucose variability is associated with increased reactive oxygen species generation, PKC activity, AGE formation, tumorigenic epigenetic changes, and apoptosis

  • Glucose variability may be associated with in vivo increased carotid artery stenosis, CVD risk, and microvascular complications of diabetes, however this data is poor

  • There is definitive lack of consensus on the in vivo data of glucose variability in diabetics, let alone nondiabetics

Our Take

Given the in vivo mechanisms of glucose variability and limited in vivo data, I perceive that elevated glucose variability is at best neutral and at worst deleterious for longevity and healthspan. Thus, Laura and I try to limit our glucose variability.

Maximum Effort (For Experts)

Reactive oxygen species generation is the main mechanism of damage in hyperglycemia

A substantial ingestion of a high glycemic index food will cause a larger blood glucose response than ingestion of low glycemic index food with equivalent macronutrient and isocaloric properties. The main mechanism of hyperglycemia-induced damage is reactive oxygen species generation (ROS) via increased flux through the mitochondrial oxidative phosphorylation pathway:

  • An in-vitro study of bovine aortic cells incubated with either 90 mg/dl glucose or 540 mg/dl sustained glucose showed ROS generation increased from 52.08 nmol/ml with 90mg/dl glucose to 154 ± 1.38 nmol/ml with 540 mg/dl glucose. This study also showed other deleterious mechanisms of hyperglycemia. When compared with 90 mg/dl of glucose, 540 mg/dl glucose caused an increase of protein kinase C from 110 pmol/min/mg to 200 pmol/min/mg and an increase in intracellular advanced glycosylation end products (AGEs) from 1.2 AU to 2.0 AU. Protein kinase C has many isoforms in the human body and its glucose-mediated damaged is associated with increases in vascular permeability and, thus, inflammation. AGEs either crosslink with other AGEs or form covalent bond with molecules to disrupt cellular function. The concentration of AGEs is elevated in both diabetics and older populations. [1]

This paper demonstrates that an increase in blood glucose levels cause many deleterious effects on the body (not just that they are correlated). Granted, 540 mg/dl is a considerable amount of blood glucose, but the question is not whether hyperglycemia causes damage, but how much and at what threshold. Additionally, blood glucose levels are never sustained at these levels in nondiabetics. Therefore, to explore this question, a new glucose parameter must be introduced: glucose variability.

Introduction of Glucose Variability

Numerous postprandial hyperglycemic events over the course of days can be characterized by glucose variability, also known as glucose excursions. Glucose variability is the measure of how far glucose strays from some baseline. For example, a scenario with a small glucose variability would be a mean average glucose of 90 mg/dl and a maximum daily glucose of 110mg/dl and minimum daily glucose of 70 mg/dl. Glucose variability has been studied since the first landmark paper was published that linked hyperglycemia to diabetic complications.

  • In 1993, the Diabetes Control and Complications Trial Research Group randomly assigned 1441 patients with insulin dependent diabetes to either intensive therapy (insulin pump or three or more insulin injections a day) or to a control group (maximum of two insulin injections a day). The results were impressive: the intensive therapy achieved an average glucose of 155 +- 30 mg/dL compared to 231 +-55 mg/dL in the control group. Outcomes demonstrated that intensive therapy decreased risk of development of retinopathy by 76%, slowed retinopathy 54%, decreased microalbuminuria 39%, albuminuria 54%, and neuropathy by 60%. This study is the foundation of the current treatment of insulin dependent diabetes. [2]

This study essentially demonstrated that hyperglycemia causes small vascular damage in diabetics and that lowering blood sugar can decrease the risk of this damage. However, not all risks were explained by HbA1c:

  • Two years later, the same group released a follow up analysis of the data which demonstrated that most of the risk of diabetic complications can be attributed to HbA1c and the amount of time that has passed since a patient was diagnosed with diabetes. There exists a percentage of risk that was not able to be controlled for by HbA1c, time since diagnosis, and twenty other covariates: approximately 6.3% of the risk for the control group and 2.5% for the intensive group. The authors state that “the HbA1c does not completely explain the risk of progression. There were patients who demonstrated retinopathy progression even though their HbA1c level was lower than the average within the intensive treatment group; likewise, there were patients who showed no progression after 9 years of follow-up even though their HbA1c level was far above the average in the conventional treatment group.” The authors go on to say that genetics, postprandial hyperglycemic excursions, or reactive hypoglycemia are additional mechanisms that are not captured by HbA1c that might explain the left-over risk of diabetic complications. [3]

Glucose Variability causes in vitro increase in PKC activity, oxidation, and apoptosis enzymes

These papers spawned numerous in vitro studies on glucose variability. Many of these studies have demonstrated that glucose variability can be just as damaging as sustained elevated glucose. For example, the following in vitro study concluded that high glucose variability increases apoptosis and oxidative stress more than sustained hyperglycemia alone.

  • Human umbilical cord veins were incubated at 90 mg/dl, 360 mg/dl, or alternating between those two values every day for 14 days. Every measurement demonstrated more deleterious effects with alternating low/high glucose than consistently elevated glucose. With respect to the control (90 mg/dl) group, consistently elevated vs. alternating glucose demonstrated: 350% increase in PKC activity vs. 520%, 0.38 micromol of nitrotyrosine vs 0.58 micromol (a marker of oxidation), 20 ng/ml 8OHdG vs 42 ng/ml (a marker of oxidation), 40% decrease in levels of Bcl-2 (anti-apoptotic), and two times higher levels of capase-3 activity (pro-apoptotic). These findings have a P<0.01. [4]

Glucose Variability associated with in vitro increase in maladaptive epigenetic changes and altered gene expression

Another study demonstrated pro-carcinogenic epigenetic changes and altered gene expression with transient high glycemia

  • In vivo, where aortic endothelial cells were incubated in high glucose (540 mg/dl) for 16 hours then incubated in physiologic glucose concentrations (90 mg/dl) for 6 days. There was sustained elevated genetic expression of the p65 even after 6 days of normoglycemia. P65 activation is associated with increase in cytokines such as IL-6 and tumorigenesis. This increase is caused by an increase in the production of methylglyoxal which, as discussed above, is created as a by product of ROS generation in the oxidative phosphorylation pathway during hyperglycemia. Methylglyoxal methylates the histone complex of p65, thus activating it. Inhibition of the creation of methylglyoxal stops this activation. [5]

Glucose Variability associated with in vivo carotid artery stenosis and effects on regression

Additional studies have demonstrated the deleterious effects of of increased blood glucose variability in humans, particularly in vascular disease via carotid artery stenosis:

  • Post glucose challenge glucose spikes are more strongly associated with carotid intima media thickness than fasting plasma glucose and HbA1c in prediabetics. [6]

  • Controlling postprandial hyperglycemia caused regression of carotid atherosclerosis in type 2 diabetics. Between two intervention groups, one with postprandial glucose peak of 148 +-28 mg/dl and one with 180 +- 32 mg/dl, carotid atherosclerosis regressed in 52% of the former group and 18% of the latter along with decreases in IL-6 and CRP, even with the same HbA1c. [7]

Post-meal Glucose Variability associated with CVD risk in nondiabetics

There is additional evidence that nondiabetic postprandial hyperglycemia is a risk factor for cardiovascular disease in otherwise healthy adults.

  • A 2004 meta-analysis of 38 prospective studies with n= 195,658 subjects followed over an average of 12 years demonstrated that those with the highest post-challenge blood glucose levels (150-194 mg/dL) had a relative risk of 1.19 of CVD when compared to those in the lowest post-challenge blood glucose levels (69-107 mg/dL), even when controlling for HbA1c and other CVD risk factors. [8]

This study shows that the differences in postprandial hyperglycemia in nondiabetic subjects can cause as much as a 19% increase in risk of cardiovascular disease. While not identical to the discussion at hand of postprandial hyperglycemia due to high glycemic index foods, this study lends credence to the fact that transient, elevated glucose levels can be a marker of CVD risk factor that can not be fully explained in terms of HbA1c.

Studies that demonstrated no difference

There have also been numerous studies that demonstrated no difference in risk factors with respect to glucose variability:

  • N=24 well-regulated T2DM subjects with CGMs demonstrated no difference in 24-hour urine PGF-2a (ROS byproduct) when stratified by standard deviation (SD) and mean average glucose excursion (MAGE). Median SD was 31 mg/dl and median MAGE was 85 mg/dl [9]

  • In a follow up study to the original DCCT trial, quarterly 7-point glucose profiles were either directly measured or calculated (in 16% of visits that were missing). This study only demonstrated a relationship between glucose variability and microalbuminuria when controlling for mean average glucose. Retinopathy and other microvascular complications were not statistically significant. [10]

Why might studies show no difference when a difference might exist?

There are a few other studies that demonstrate no relationship between either ROS generation or clinical outcomes and glucose variability. However, there are two main confounders that make researching glucose variability difficult:

  1. Fidelity of data. Glucometer testing, even if it is a 7-point test (ie, the subject measures their blood glucose at 7 predefined points during the day, such as fasting, post-breakfast, pre-lunch, etc) cannot accurately map glucose variation throughout the day. Additionally, from personal experience and market research, current CGMs are approved by the FDA to be as inaccurate as 20% of the real blood glucose value. These two points add noise to a probably small signal, thus increasing the risk of type-II errors.

  2. Characterization of data. There is no current consensus on the best measurement of glucose variability. Standard deviation and MAGE are the most common, although there have been dozens measured. It is unknown which of these accurately captures glucose variability the best.

Final Thought

Normal, nondiabetic postprandial hyperglycemia has peaked for Laura and I at around 200 (see our CGM data page). While not close to the 350-500 mg/dl seen in these studies and others, we feel there is benefit to decreasing the flux through the oxidative phosphorylation pathway in order to decrease ROS generation. We already know that caloric restriction is the best studied way to increase lifespan. An overall decrease in ROS generation, along with decreasing the other deleterious effects of hyperglycemia, is a feasible mechanism for at least part of this lifespan increase.

[1] Normalizing mitochondrial superoxide production blocks three pathways of hyperglycaemic damage. Takeshi Nishikawa et al., Nature 404, 787-790 (2000)

[2] The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. DCCM, NEJM 1993 329(14):977-86

[3] Association of Glycemic Variability in Type 1 Diabetes With Progression of Microvascular Outcomes in the Diabetes Control and COmplications Trial. Lachin JM et al., Diabetes Care 2017 40(6):777-783

[4] Intermittent High Glucose Enhances Apoptosis Related to Oxidative Stress in Human Umbilical Vein Endothelial Cells. L Quagliaro et al., Diabetes 2003 Nov;52(11) 2795-2804

[5] Transient high glucose causes persistent epigentic changes and altered gene expression during subsequent normoglycemia. A El-Osta et al., JEM 2008 205 (10):2409

[6] Postchallenge plasma glucose and glycemic spikes are more strongly associated with atherosclerosis than fasting glucose or HbA1c level. TS Temelkova-Kurktschiev et al., Diabetes Care 2000 23(12):1830-1834

[7] Regression of Carotid Atherosclerosis by Control of Postprandial Hyperglycemia in Type 2 Diabetes Mellitus. K Esposito et al., Circulation. 2004;110:214-219

[8] Is Nondiabetic Hyperglycemia a Risk Factor for Cardiovascular Disease? E Levitan et al., Arch Intern Med. 2004;164(19):2147-2155

[9] No Relevant Relationship between Glucose Variability and Oxidative Stress in Well-Regulated Type 2 Diabetes Patients. S Siegelaar et al., Diabetes Science and Technology 2011 5(1)86-92

[10] Association of Glycemic Variability in Type 1 Diabetes With Progression of Microvascular Outcomes in the Diabetes Control and Complications Trial. Lachin JM et al., Diabetes Care. 2017 Jun;40(6)777-783

bottom of page