GH-METHODS

Math-Physical Medicine

NO. 390

Applying the concept of glycemic variability (glucose fluctuation) for an extended study on the self-recovery of pancreatic beta cells and risk probability of having a cardiovascular disease or stroke using GH-Method: math-physical medicine

Corresponding Author: Gerald C. Hsu, eclaireMD Foundation, USA.

Abstract
The author believes that any newly created biomarker should accurately describe the biomedical phenomena of a disease but also be easy enough for physicians, even patients, to comprehend and be applied in their disease control.

The concept and discussion of glycemic variability (GV) have existed more than a decade based on the clinical usage and results of the continuous glucose monitoring (CGM) devices to monitor severe diabetes patients and insulin treatments in hospitals.  Therefore, thus far, most of reports the author has read were based on 2 to 3 days collected glucose data.  Although many GV medical papers have been published, but unfortunately only very few of them have provided realistic connections to the biomedical intuitive understanding along with the commonly accepted and useful equation for its general applications on diabetes control. Therefore, GV remains a research subject floating in the medical research community, instead of being used as a common clinical tool for both diabetes patients and their physicians.  The author has had type 2 diabetes (T2D) for 25 years and is under the care of physicians associated with a renown medical school in the US.  While he has met numerous clinical doctors in the past decade, GV related terminologies have never been mentioned.

Defining GV remains a challenge primarily due to the difficulty of data collection with its associated necessary tasks, such as data transfer, clean, process, and analysis which resulted into the ambiguity of GV’s existing interpretations, along with the lack of consensus regarding the optimal approach for its clinical management.  For example, one of the major GV derivations, mean amplitude of glycemic excursion (MAGE), involves the usage of standard deviation (SD) from statistics.  Although SD is widely used, it has some limitations since its inherited assumption of measured glucose data are normally distributed, which is typically not the case for most glucose waves.  The author is a mathematician, physicist, and engineer who has collected his own 2,958 postprandial plasma glucose (PPG) waves with 986 daily glucose waves.  He has further used various analytical approaches to analyze his data and waveforms from many different research angles.  Therefore, he thoroughly understands the behaviors and characteristics of many glucose waves from his own body.

Over the past decade, many medical research scientists have used glucose data within a time span of 2 to 3 days collected from patients in hospitals or clinic centers, to conduct their research or clinical emergency treatment plans.  On the contrary, by taking the glucose data over a much longer time span for several years from diabetes outpatients via a CGM device only became available after 2017.  Of course, a year long wave dataset would provide much more biomedical information than a day long wave dataset.  The other mundane tasks of using the self-monitored blood glucose (SMBG) devices include utilizing the electronic devices to automatically transfer large amounts of glucose data (e.g., 288 data per day at every 5 minutes) to a computer or iPhone instead of time-consuming manually data recording and transferring, and then performing the necessary follow-on technical tasks of data cleaning, data processing, data management, and data analysis which require proper training and sufficient knowledge in the domain of computer science, mathematics, and physics.  All of the above-mentioned tasks are still challenging for most diabetes patients, physicians, and even some of medical research scientists.  Due to lack of professional training and academic knowledge in this domain, most patients and clinical physicians have encountered difficulties in understanding and applying the GV related materials.

Data without careful cleaning and proper preparation belong to the category of “garbage inputs” which would result in “garbage outputs”.

Based on the above-mentioned theoretical and technical viewpoints, the author decided to conduct this study on applying the basic concept of glycemic variability (i.e., glucose fluctuation between peak and trough) in combination with the primary characteristics of wave theory (mainly frequencies, amplitudes, phases, and associated energies of glucoses).  This would assist in his investigation of the self-recovery of his pancreatic beta cells and various internal organ impacts from the energy associated with glucose waves, including GV (glucose fluctuations).  In this article, the author applies the ratios of the glucose fluctuation (defined as using glucose peak as the maximum glucose minus the glucose trough as the minimum glucose and then divided by the average glucose), in combination with the mean glucose values (HbA1C) and energy associated with glucose fluctuation using Fourier Transform into a  frequency domain of glucose energy to investigate the self-recovery of his pancreatic beta cells and his risk probabilities of having a cardiovascular disease (CVD) or stroke.

In addition to the postprandial beta cell dysfunction (health of the pancreatic insulin), the GV concept may provide useful clues on the increased rate of diabetic related complications, such as both macro- and micro-blood vessels related diseases, including CVD and stroke. Finally, the GV concept may also offer additional assistance with the reduction of the mortality rate for hypoglycemic patients in intensive care units.

Although GV can be used as an indicator for insulin resistance, diabetes complications, and hypoglycemic risk for ICU patients, the author focuses on his continuous medical research work for the “self-recovery” of his pancreatic beta cells.  He uses “self-recovery” because he has kept his carbs/sugar intake amount less than 15 grams per meal and his post-meal walking exercise more than 4,000 steps over the past 5 years.  Since 12/8/2015, he has also ceased taking any diabetes medication, which is the strongest influential factor for the phenomena of glucose fluctuations. Therefore, his body is totally free of any external chemical intervention that may alter the internal organ’s biochemical process and reactions. Under this strict controlled lifestyle and environment, his damaged pancreatic beta cells must go through the self-repairing process in order to show any meaningful improvement signs of his diabetes conditions.  This is his chosen approach in “fixing his diabetes conditions from their root causes via a stringent lifestyle management”.

In this article, he utilized four different but still inter-related analysis approaches.  The first approach is to evaluate the glucose fluctuation difference from his daily CGM collected glucose data, which is defined as the maximum glucose minus the minimum glucose and then divided by the average glucose.  The second approach is to calculate and evaluate his specific percentages of hypoglycemia and hyperglycemia. The third approach is to assess his pancreatic beta cells condition via the traditional HbA1C value changes, which is the mean value of glucoses, not the short-term or long-term glucose fluctuations.  The fourth approach is to utilize both energy and wave theories and his developed metabolism model to evaluate the energy associated with both glucose and glycemic fluctuation to compare his risk probabilities of having a CVD or stroke between Y2019 and Y2020.   All of the four analysis approaches have demonstrated that his COVID-19 Period B performance in Y2020 is better than his non-virus Period A performance in Y2019. However, the category of his glucose associated energy and metabolism-based calculation of CVD risk have lower performance percentages (7% to 9%) than the category of his glucose and glycemic fluctuation (13% to 28%).

Introduction
The author believes that any newly created biomarker should accurately describe the biomedical phenomena of a disease but also be easy enough for physicians, even patients, to comprehend and be applied in their disease control.

The concept and discussion of glycemic variability (GV) have existed more than a decade based on the clinical usage and results of the continuous glucose monitoring (CGM) devices to monitor severe diabetes patients and insulin treatments in hospitals.  Therefore, thus far, most of reports the author has read were based on 2 to 3 days collected glucose data.  Although many GV medical papers have been published, but unfortunately only very few of them have provided realistic connections to the biomedical intuitive understanding along with the commonly accepted and useful equation for its general applications on diabetes control. Therefore, GV remains a research subject floating in the medical research community, instead of being used as a common clinical tool for both diabetes patients and their physicians.  The author has had type 2 diabetes (T2D) for 25 years and is under the care of physicians associated with a renown medical school in the US.  While he has met numerous clinical doctors in the past decade, GV related terminologies have never been mentioned.

Defining GV remains a challenge primarily due to the difficulty of data collection with its associated necessary tasks, such as data transfer, clean, process, and analysis which resulted into the ambiguity of GV’s existing interpretations, along with the lack of consensus regarding the optimal approach for its clinical management.  For example, one of the major GV derivations, mean amplitude of glycemic excursion (MAGE), involves the usage of standard deviation (SD) from statistics.  Although SD is widely used, it has some limitations since its inherited assumption of measured glucose data are normally distributed, which is typically not the case for most glucose waves.  The author is a mathematician, physicist, and engineer who has collected his own 2,958 postprandial plasma glucose (PPG) waves with 986 daily glucose waves.  He has further used various analytical approaches to analyze his data and waveforms from many different research angles.  Therefore, he thoroughly understands the behaviors and characteristics of many glucose waves from his own body.

Over the past decade, many medical research scientists have used glucose data within a time span of 2 to 3 days collected from patients in hospitals or clinic centers, to conduct their research or clinical emergency treatment plans.  On the contrary, by taking the glucose data over a much longer time span for several years from diabetes outpatients via a CGM device only became available after 2017.  Of course, a year long wave dataset would provide much more biomedical information than a day long wave dataset.  The other mundane tasks of using the self-monitored blood glucose (SMBG) devices include utilizing the electronic devices to automatically transfer large amounts of glucose data (e.g., 288 data per day at every 5 minutes) to a computer or iPhone instead of time-consuming manually data recording and transferring, and then performing the necessary follow-on technical tasks of data cleaning, data processing, data management, and data analysis which require proper training and sufficient knowledge in the domain of computer science, mathematics, and physics.  All of the above-mentioned tasks are still challenging for most diabetes patients, physicians, and even some of medical research scientists.  Due to lack of professional training and academic knowledge in this domain, most patients and clinical physicians have encountered difficulties in understanding and applying the GV related materials.

Data without careful cleaning and proper preparation belong to the category of “garbage inputs” which would result in “garbage outputs”.

Based on the above-mentioned theoretical and technical viewpoints, the author decided to conduct this study on applying the basic concept of glycemic variability (i.e., glucose fluctuation between peak and trough) in combination with the primary characteristics of wave theory (mainly frequencies, amplitudes, phases, and associated energies of glucoses).  This would assist in his investigation of the self-recovery of his pancreatic beta cells and various internal organ impacts from the energy associated with glucose waves, including GV (glucose fluctuations).  In this article, the author applies the ratios of the glucose fluctuation (defined as using glucose peak as the maximum glucose minus the glucose trough as the minimum glucose and then divided by the average glucose), in combination with the mean glucose values (HbA1C) and energy associated with glucose fluctuation using Fourier Transform into a  frequency domain of glucose energy to investigate the self-recovery of his pancreatic beta cells and his risk probabilities of having a cardiovascular disease (CVD) or stroke.

In addition to the postprandial beta cell dysfunction (health of the pancreatic insulin), the GV concept may provide useful clues on the increased rate of diabetic related complications, such as both macro- and micro-blood vessels related diseases, including CVD and stroke. Finally, the GV concept may also offer additional assistance with the reduction of the mortality rate for hypoglycemic patients in intensive care units.

Methods
1. MPM Background:
To learn more about the author’s GH-Method: math-physical medicine (MPM) methodology, readers can refer to his articles to understand his developed MPM methodology in References 1 and 2.

2. Other GV research work:
There are many available articles regarding GV; however, the author decided to combine five published articles into one excerpt (References 3, 4, 5, 6, and 7).  These 5 references have cited a total of 201 published papers.  This allow readers not having to search for some key information from a long list of the cited reference articles.

References 3 concentrates on the comparison of many published GV articles.  Reference 4 focuses on an algorithm, method, and firmware design of a web-based APP software in calculating GV values.  Reference 5 evaluates the relationship between glycemic variability and pancreatic beta-cell dysfunction.  Reference 6 from the American Diabetes Association (ADA) describes the overall picture of GV.  Reference 7 defines the mathematical equation of MAGE.

Here is the combined excerpt:

Several pathophysiological mechanisms were reported, unifying the two primary mechanisms: excessive protein glycation end products and activation of oxidative stress, which causes vascular complications.  Intermittent high blood glucose exposure, rather than constant exposure to high blood glucose, has been shown to have deleterious effects in experimental studies.  In in-vitro experimental settings and in animal studies, glycemic fluctuations display a more deleterious effect on the parameters of CV risk, such as endothelial dysfunction.  There is a significant association between GV and the increased incidence of hypoglycemia. Hypoglycemic events may trigger inflammation by inducing the release of inflammatory cytokines.  Hypoglycemia also induces increased platelet and neutrophil activation. The sympathoadrenal response during hypoglycemia increases adrenaline secretion and may induce arrhythmias and increase the cardiac workload. Underlying endothelial dysfunction leading to decreased vasodilation may contribute to CV risk.  Published studies have demonstrated that GV, particularly when associated with severe hypoglycemia, could be harmful not only to people with diabetes but also to non-diabetic patients in critical care settings.  Overall, the pathophysiological evidence appears to be highly suggestive of GV being an important key determinant of vascular damage.  In addition to HbA1c, GV may have a predictive value for the development of T1DM complications.  In insulin-treated T2DM, the relevance of GV varies according to the heterogeneity of the disease, the presence of residual insulin secretion and insulin resistance.  HbA1c is a poor predictor of hypoglycemic episodes because it only considers 8% of the likelihood of severe hypoglycemia; on the contrary, GV can account for an estimated 40% to 50% of future hypoglycemic episodes.  HbA1c is a poor predictor of hypoglycemic risk, whereas GV is a strong predictor of hypoglycemic episodes. GV was an independent predictor of chronic diabetic complications, in addition to HbA1c. We should note that PPG and GV are not identical, even if they are closely related.  The attention dedicated to GV is derived from the above evidence concerning its effects on oxidative stress and, from the latter, on chronic diabetes complications.  Control of GV has been the focus of a number of interventional studies aimed at reducing this fluctuation. Diet and weight reduction are the first therapeutic instrument that can be used for reducing GV.  

Despite the various formulas offered, simple and standard clinical tools for defining GV have yet to evolve and different indexes of GV should be used, depending on the metabolic profile of the studied population.  Moreover, the absence of a uniformly accepted standard of how to estimate postprandial hyperglycemia and GV adds another challenge to this debate.

The majority of these studies have used time-averaged glucose values measured as glycosylated hemoglobin (HbA1c), an indicator of the degree of glycemic control, which is why HbA1c has become the reference parameter for therapies aimed at reducing the risk of complications from diabetes. Chronic hyperglycemia is almost universally assessed by HbA1c which has been shown to correlate closely with mean glucose levels over time, as determined by continuous glucose monitoring (CGM).  However, the relative contribution of postprandial glycemic excursions and fasting to overall hyperglycemia has been the subject of considerable debate. Monnier et al. suggested that the relative contributions of fasting and postprandial glucose differ according to the level of overall glycemic control.  Fasting glucose concentrations present the most important contribution to hemoglobin glycosylation, whereas at lower levels of HbA1c, the relative contribution of postprandial hyperglycemia becomes predominant. Collectively, GV is likely to be incompletely expressed by HbA1c, particularly in patients with good metabolic control.  

GV is a physiological phenomenon that assumes an even more important dimension in the presence of diabetes because it not only contributes to increasing the mean blood glucose values but it also favors the development of chronic diabetes complications. It appears that GV is poised to become a future target parameter for optimum glycemic control over and above standard glycemic parameters, such as blood glucose and HbA1c.  Avoiding both hyperglycemia and hypoglycemia by careful use of SMBG and the availability of new agents to correct hyperglycemia without inducing hypoglycemia is expected to reduce the burden of premature mortality and disabling CV events associated with diabetes mellitus.  However, defining GV remains a challenge primarily due to the difficulty of measuring it and the lack of consensus regarding the most optimal approach for patient management.

The risk of developing diabetes-related complications is related not only to long-term glycemic variability, but may also be related to short-term glucose variability from peaks to nadirs.  Oscillating glucose concentration may exert more deleterious effects than sustained chronic hyperglycemia on endothelial function and oxidative stress, two key players in the development and progression of cardiovascular diseases in diabetes. Percentages of hyperglycemia (levels between 126 and 180 mg/dl) and hypoglycemia (levels below 70.2 mg/dl) episodes should be used in the GV related research.  Mean amplitude of glycemic excursions (MAGE), together with mean and SD, is the most popular parameter for assessing glycemic variability and is calculated based on the arithmetic mean of differences between consecutive peaks and nadirs of differences greater than one SD of mean glycemia. It is designed to assess major glucose swings and exclude minor ones.  

The features discouraging use of glycemic variability as a parameter in clinical practice and trials are the difficulty of interpreting numerous parameters describing this phenomenon and a limited number of computational opportunities allowing rapid calculation of glycemic variability parameters in CGM data.

The UK Prospective Diabetes Study (UKPDS) showed that after an initial improvement, glycemic control continues to deteriorate despite the use of oral agents to enhance insulin secretion and to reduce insulin resistance.  This deterioration can be attributed to the progressive decline of β-cell function.  Even in subjects with well-controlled type 2 diabetes, 70% of the variability of A1C can be explained by abnormalities in postprandial glucose.  Chronic sustained hyperglycemia has been shown to exert deleterious effects on the β-cells and the vascular endothelium. In vivo studies have convincingly demonstrated that hyperglycemic spikes induce increased production of free radicals and various mediators of inflammation, leading to dysfunction of both the vascular endothelium (3) and the pancreatic β-cell.

3. Mean amplitude of glycemic excursions (MAGE):
Furthermore, such a measure should be simple in concept and faithful to the physiological basis for the glucose swings.  Because interest lay in the amplitude of glycemic swings and not in the dispersion of all the glucose data, SD was considered to be unsuitable.  The criterion, which did recognize all of the meal-related glucose excursions for all of the normal subjects, was the SD of the mean BG for each 24-h period of study (288 values taken q5min from the continuous record) for each individual. In contrast, 0.5 SD and 1.5 SD were less inclusive/exclusive. Although the numerical value of 1 SD will perforce differ in absolute value from person to person, it nevertheless acts as an individualized standard. By convention, a glycemic excursion (both trough-to-peak and peak-to-trough) must exceed 1 SD of the respective 24-h BG profile. For continuous recordings exceeding 24 h, the use of 1 SD calculated for the whole period of study may result in the inclusion of the same excursions as use of the separate 24-h SDs, since SDs from successive days do not differ by much (even in type 1 diabetic patients as long as therapy has not changed during the period of monitoring).  Only one limb of the excursion, ascending or descending, determined by the initial excursion (which is not always an inflection especially in type 1 diabetic patients) is used for calculation of subsequent excursions.  Should the subsumed excursion be of a magnitude observed for normal subjects its exclusion may be inconsequential relevant to the risk for the development of microvascular complications of diabetes.  The arithmetic mean of the glycemic excursions for the period of study (24 h, 48 h, or longer) is the value of mean amplitude of glycemic excursions (MAGE).  An automated algorithm has been created for the calculation of MAGE.  Although created for determination from continuous BG analysis, MAGE has been applied to intermittent (7- and 22-point sampling/24 h) measurements as well as continuous interstitial glucose monitoring.  

However, the biomedical intuitive feeling from above equation is not as clear, strong, or direct as the following simplified formula defined by the author:

Glucose fluctuation rate = [Summation (i = 1 to n) of : ((maximum glucose – minimum glucose) / average glucose)] / n

4. Data and analysis in this study:
The author has collected 96 glucose data per day (ATH every 15 minutes) using a CGM device since 5/5/2018 and 288 glucose data collected per day (at every 5 minutes) since 2/17/2020.  During the past 971 days (5/5/2018 – 12/31/2020), he has collected 93,216 glucose data (15-minute model).  He decided to use this 15-minute model due to its sufficient long period of data collection.  For a major part of this study, in order to compare equal-length data and results, he divided the collected data from the 15-minute model into two equal-length time periods based on his distinguished lifestyles: Period A for 2019 (from 1/1/2019 to 12/31/2019) with heavy traveling schedules and busy lifestyle along with Period B for 2020 (from 1/1/2020 – 12/31/2020) with a calm, peaceful, non-traveling quarantine lifestyle.

In this article, he chose not to use the defined MAGE term and equation due to its controversial arguments and comprehension difficulty by both patients and physicians as cited in above excerpt.  Instead, his research work was conducted by using four different but inter-related key components which are easier for both patients and physician to understand. The first component is the glucose fluctuation rate, which is defined as the glucose difference between maximum glucose and minimum glucose divided by average glucose. The second component is the occurrence frequency of hypoglycemia (less than 70 mg/dL) and hyperglycemia (both greater than 140 mg/dL and greater than 180 mg/dL), which is expressed in terms of the occurrence percentages of hypoglycemia and hyperglycemia in terms of total glucose data amount.  The third component is a routinely utilized value by clinical physicians and diabetes patients, i.e. the changes of HbA1C values.  The HbA1C is related to the average glucose value (the mean value) which is capable to provide a good overall picture of most diabetes conditions.  The forth component is the risk probability percentage of having a CVD or stroke using his previously developed metabolism model.

Mathematically, the normal distribution’s characteristic function is defined by just two moments: mean and variance (or standard deviation, the “SD”).  The mean tells you where the center of your distribution is, while the standard deviation tells you how close to the center your data is.  Therefore, for normal distribution, the standard deviation is especially important, because it contributes about 50% of these two definitions or moments.  However, for other data distribution types, the standard deviation is in some ways less important because non-normal distribution types contain other moments.  We cannot assume that most of our data (business, social, economic, scientific or medical origin) are approximately “Normal”, i.e., they are generated by a Gaussian process or by a sum of multiple such processes.  Therefore, when we define a new biomarker with its mathematical operations heavily depending on SD, we must be aware of the limitation of the biomarker’s accuracy and applicability due to its non-Gaussian distribution type in real life.

Let us refer to the original concept of glycemic variability (GV) which attempts to find the deleterious effects on the internal organs due to the glucose fluctuations.  Glucose fluctuations dealing with a glucose wave is not a normal Gaussian type of wave, which has fluctuations within both longer timespan (several years) and shorter timespan (several days).  The most important things we are trying to explore and discover are the frequency of occurrence (how often the fluctuations happen) and amplitude variation (how violently the fluctuations are between the peak and trough or nadir, i.e., maximum value minus minimum value) of these glucose waves.  Therefore, there is no need to apply such a complex mathematical equation of the Mean Amplitude of Glycemic Excursion (MAGE) as defined thus far.  This MAGE involves mathematical operations such as mean, logarithm, and standard deviation which is difficult for both patients and physicians to understand.  More importantly, during the complex definition and multiple mathematical operations, it has lost its direct linkage with an easy-to-understand biomedical intuitions to users.  That is why the MAGE remains as a terminology in medical research arena only.  In the author’s opinion, once physicians or patients can understand, feel, and then control their “glucose fluctuation” situations, either within a few days or a few years (of course, the longer period of time the better), their objective of diabetes control can then be met and not have to deal with a complicated mathematical equation for biomarker such as MAGE.

5. Candlesticks (K-Line) Model:
A Japanese merchant, who traded in the rice market in Osaka, Japan, started the candlestick charting around 1850.  An American fellow, Steve Nison, brought the candlestick model concept and method to the Western world in 1991.  These techniques are largely used in today’s stock market to predict the stock price trend.

The author was the CEO of a public-traded company.  Therefore, he is quite familiar with the Candlestick model or the K-line model.  On 4/17/2018, he had an idea to study glucose behavior by using the candlestick chart and subsequently developed a customized software to analyze his big data of glucose.  The analogies between fluctuations of stock price and glucose value are described as follows:

  • Stock prices are closely related to the psychology of the buyers and sellers, which is similar to the glucoses related to a patient body’s biochemical interactions and lifestyle behaviors.
  • Stock price wave of a public traded company is dependent upon its product line, internal management, marketing efforts, and public events and customer perception.  This is remarkably similar to the PPG wave of a diabetes patient being dependent on his/her complex food & diet (buying stock), exercise pattern and amount (selling stock), weather temperature (buying and selling stock), and pancreatic beta cell insulin function (similar to SEC regulations).  From a trained mathematician’s eyes, these biomedical wave and financial wave are just two different but similar mathematical representations.  Wave theory can be applied on both of their behaviors.
  • When there are more buyers than sellers, the stick price goes up, which is similar to the glucose value rising when carbs/sugar intake increases which infuses the energy generation (more buyers) or lack of exercise which reduces the energy consumption (less sellers).
  • When there are more sellers than buyers, price goes down, which is similar to the glucose value decreasing when carbs/sugar intake decreases (less buyers) or exercise increases (more sellers).

During his period of using the CGM sensor to collect his glucoses data, his standard PPG wave covers a period of 180 minutes, or 3 hours from the first bite of his meal.  Each PPG waveform contains the following five key characteristic data:

  1. “Open” value as his PPG at first-bite, 0 minute
  2. “Close” value as PPG at 180 minutes
  3. “Minimum” value as the lowest PPG
  4. “Maximum” value as the highest PPG
  5. “Average” glucose (HbA1C) as the average value of 13 recorded PPG data per meal over 3 hours

6. Metabolism model:
In 2014, the author applied a topology concept of mathematics and finite-element method of engineering, to develop a ten-dimensional complex mathematical model of metabolism which contains four output categories (weight, glucose, BP, and lipids) and other lab-tested data (ACR, TSH, and others), and six input categories (food, water intake, exercise, sleep, stress, and routine life patterns), and in total, about 500 detailed elements. He has further defined two new parameters, metabolism index (MI), as the combined score of the above 10 metabolism categories (dimensions) and 500 detailed elements, and general health status unit (GHSU), as the 90-days moving average value of MI.  Please noted that Mi (where i = 1 through 10) represents individual metabolism score of each category.   Since 2012, he has collected ~2 million data of his own biomedical conditions and personal lifestyle details.  He utilized this sophisticated MI model to calculate his risk probabilities of having CVD, stroke, diabetic kidney disease (DKD), and even ~40% of cancers.  Of course, the GV is only a small part of the influential factors of this metabolism model.  However, all of metabolism inputs and outputs are inter-related, which is similar to our internal organs.  Therefore, the risk of having the metabolic induced chronic disease complications are closely related to the total energy generated by excessive amount of glucose (particularly hyperglycemia) and the situations of glucose fluctuations, i.e. glycemic variability.

Results
Figure 1 shows glucose fluctuation percentage (maximum glucose minus minimum glucose and then divided by the average glucose) of 61%, average CGM sensor glucoses of 131 mg/dL, and their waveform comparison with correlation coefficient of 10% (no correlation) for the non-virus Period A (Y2019).

Figure 1: Glucose fluctuation vs. Sensor glucose (Non-virus Y2019)

Figure 2 depicts glucose fluctuation percentage of 56%, average CGM sensor glucoses of 119 mg/dL, and their waveform comparison with high correlation coefficient of 63% (strong positive correlation) for the COVID-19 Period B (Y2020). 

The following table lists the data comparison in a format of Y2019 vs. Y2020:

  • Glucose fluctuation %: 61%, 56%
  • Average sensor glucose: 131, 119
  • Correlation coefficient: 10%, 63%

There are two interesting findings from comparing Figure 1 and Figure 2.  The first is that both glucose fluctuations and average sensor glucoses of the COVID-19 Period B is lower than the non-virus Period A.  The second is that Period A has an extremely low correlation of 10% (almost none) between glucose fluctuation and glucose itself, while Period B has a strong correlation of 63% between glucose fluctuation and glucose itself.  

Figure 2: Glucose fluctuation vs. Sensor glucose (COVID-19 Y2020)

Figure 3 indicates a direct comparison of the collected daily glucose fluctuation percentage data between Period A of 62% and Period B of 55%.  These two percentages have a 1% difference from the glucose fluctuation percentages from the results in Figures 1 and 2.  This is based on the 90-days moving average data that provides a better graphical view of the waveform moving trends, while Figure 3 uses the daily data for accuracy of glucose fluctuation computation.

Figure 3: Sensor daily glucose fluctuation ((max-min)/avg) for both non-virus period (62%) and COVID-19 period (55%)

Figure 4 illustrates a bar diagram to demonstrate the conclusive results of this particular study.  A table lists the results in the format of Y2020, Y2019, and % of Y2019 divided by Y2020 which are displayed as follows:

  • Average glucose: 116 / 131 / 113%
  • Max-Min glucose: 64 /   81 / 128%
  • (Max-Min)/Avg: 55%/ 62%/ 113%
  • Glucose Energy: 33 / 35 / 107%
  • CVD/Stroke Risk:  52%/57% /109%

In summary, Y2019 has glucose and fluctuation being 13% higher than Y2020, along with the maximum glucose minus minimum glucose being 28% higher than Y2020.  At least, these differences are within a similar bracket of 13% to 28%; however, Y2019 have glucose associated energy being 7% higher than Y2020 with CVD/stroke risk 9% higher than Y2020, which are within a lower bracket of 7% to 9%.  

Figure 4: Comparison of Glucose, Fluctuation, (max-min)/avg, glucose associated energy, and CVD/Stroke risk probability

Figure 5 reflects the comparison between 90-days moving average of daily glucose (both finger measured and CGM sensor collected) and HbA1C values for Y2019 and Y2020.  In summary, all the four correlations are extremely high, between 81% to 99%, which shows HbA1C is correlated to the mean values of daily glucoses, regardless of measuring devices.  However, HbA1C and the mean glucoses are still extremely useful for both diabetes patients and physicians to obtain an overview of glucose situations and diabetes control.  Of course, GV and glucose fluctuations can provide additional information about diabetes disease and its complications, such as organ impact due to excessive glucoses, but any new biomarker must be easily understood and applied to diabetes control.  That is why the author selected to apply the GV concept of glycemic fluctuation only instead of the defined equation of MAGE.  

Figure 5: Sensor glucose vs. HbA1C comparison

Figure 6 reveals the percentages of both hypoglycemic case (<70 mg/dL) and hyperglycemic cases (both>140 mg/dL and >180 mg/dL) for these two periods.  The data table, in the lower portion of Figure 6, signifies these percentages while the bar chart, in the upper portion, provides a clear graphic view of these percentage differences.  The hypoglycemic situations with almost 0% indicate a low risk of having an insulin shock for the author.  On the other hand, for the hyperglycemic situations of both >140 mg/dL and >180 mg/dL, the COVID-19 Period B have consistent lower percentages than the non-virus Period A.  Again, this Figure 6 also indicates his observed improvement or self-recovery of his pancreatic beta cells.  

Figure 6: Comparison of percentages of Hypoglycemia (<70) and Hyperlipidemia (>140 & >180)

Conclusions
Although GV can be used as an indicator for insulin resistance, diabetes complications, and hypoglycemic risk for ICU patients, the author focuses on his continuous medical research work for the “self-recovery” of his pancreatic beta cells.  He uses “self-recovery” because he has kept his carbs/sugar intake amount less than 15 grams per meal and his post-meal walking exercise more than 4,000 steps over the past 5 years.  Since 12/8/2015, he has also ceased taking any diabetes medication, which is the strongest influential factor for the phenomena of glucose fluctuations. Therefore, his body is totally free of any external chemical intervention that may alter the internal organ’s biochemical process and reactions. Under this strict controlled lifestyle and environment, his damaged pancreatic beta cells must go through the self-repairing process in order to show any meaningful improvement signs of his diabetes conditions.  This is his chosen approach in “fixing his diabetes conditions from their root causes via a stringent lifestyle management”.

In this article, he utilized four different but still inter-related analysis approaches.  The first approach is to evaluate the glucose fluctuation difference from his daily CGM collected glucose data, which is defined as the maximum glucose minus the minimum glucose and then divided by the average glucose.  The second approach is to calculate and evaluate his specific percentages of hypoglycemia and hyperglycemia. The third approach is to assess his pancreatic beta cells condition via the traditional HbA1C value changes, which is the mean value of glucoses, not the short-term or long-term glucose fluctuations.  The fourth approach is to utilize both energy and wave theories and his developed metabolism model to evaluate the energy associated with both glucose and glycemic fluctuation to compare his risk probabilities of having a CVD or stroke between Y2019 and Y2020.   All of the four analysis approaches have demonstrated that his COVID-19 Period B performance in Y2020 is better than his non-virus Period A performance in Y2019. However, the category of his glucose associated energy and metabolism-based calculation of CVD risk have lower performance percentages (7% to 9%) than the category of his glucose and glycemic fluctuation (13% to 28%).

References

  1. Hsu, Gerald C., eclaireMD Foundation, USA, “Biomedical research using GH-Method: math-physical medicine, version 3 (No. 386)”
  2. Hsu, Gerald C., eclaireMD Foundation, USA, “From biochemical medicine to math-physical medicine in controlling type 2 diabetes and its complications (No. 387)”
  3. Sunghwan Suh and Jae Hyeon Kim, “Glycemic Variability: How Do We Measure It and Why Is It Important?”; Diabetes & Metabolism Journal, https://www.ncbi.nlm.nih.gov › pmc; http://www.e-dmj.com
  4. Dorota Czerwoniuk, M.Sc., Wojciech Fendler, M.D., Lukasz Walenciak, M.Sc., and Wojciech Mlynarski, M.D., Ph.D.; Department of Pediatrics, Oncology, Hematology, and Diabetology, Medical University of Lodz, Lodz, Poland, Journal of Diabetes Sci Technol. 2011 Mar; 5(2): 447–451, Published online 2011 Mar 1; doi:1177/193229681100500236, PMCID: PMC3125941, PMID: 21527118
  5. Klaus-Dieter Kohnert, MD, PHD, Petra Augstein, MD, PHD, Eckhard Zander, MD, Peter Heinke, MSC, Karolina Peterson, MD, Ernst-Joachim Freyse, MD, PHD, Roman Hovorka, PHD3 and Eckhard Salzsieder, PHD1; Author Affiliations:1Institute of Diabetes “Gerhardt Katsch,” Karlsburg, Germany; 2Clinics for Diabetes and Metabolic Diseases, Karlsburg, Germany; 3Institute of Metabolic Science, University of Cambridge, Cambridge, U.K.; “Glycemic Variability Correlates Strongly With Postprandial β-Cell Dysfunction in a Segment of Type 2 Diabetic Patients Using Oral Hypoglycemic Agents”
  6. John Service; Glucose Variability; ADA, Diabetes 2013 May; 62(5): 1398-1404. https://doi.org/10.2337/db12-1396; service.john@mayo.edu.
  7. Xuefei Yu, Liangzhuo Lin, Jie Shen,Zhi Chen,Jun Jian, Bin Li,1 and Sherman Xuegang Xin; Calculating the Mean Amplitude of Glycemic Excursions from Continuous Glucose Data Using an Open-Code Programmable Algorithm Based on the Integer Nonlinear Method; Computational and Mathematical Methods in Medicine / 2018 / Volume 2018 |Article ID 6286893 | https://doi.org/10.1155/2018/6286893;