Math-Physical Medicine

NO. 012

Relationship Between Glucose and Blood Pressure

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

Since 1997, the author has been diagnosed with three chronic diseases such as type 2 diabetes (T2D), hypertension, and hyperlipidemia.  From 2012, he has maintained a disciplined lifestyle program, while collecting ~1.5M data regarding his own health and body conditions.  He later developed a mathematical model of the human metabolism system, which contains 11 categories and ~500 elements.  For example, in this model, blood pressure conditions are under the category M3 (Metabolism Index 3).  This paper discusses specifically the relationship between two categories of health conditions: glucose and blood pressure (BP). 

Using the finger-piercing method, the author, who is also the patient, measures his fasting plasma glucose (FPG) in the early morning before starting his meals and activities.  He also measures his postprandial plasma glucose (PPG) three times a day, approximately two hours after the first-bite of each meal.  He measures his BP at least once a day, preferably in the early morning, but sometimes multiple measurements as needed.

He collects various detailed lifestyle data such as medication, salt intake, stress, sleep, illness, water, carbs/sugar intake, exercise, weight, weather, etc.  As a result, he has utilized these big data to conduct many useful correlation and interrelationship studies between chronic diseases and their contributing lifestyle factors.

Initially, he defined two simple equations for further calculations:

  1. Daily average glucose mg/dL = (1 FPG+3 PPG)/4
  2. Daily average blood pressure (M3: Metabolism Index 3) = 1.0+((SBP-120)/120+(DBP-80)/80)/2

It should be noted here that the heart rate is not included in his equation.

The detailed comparison results can be displayed using time-series analysis of two sets of data in a form of digitized curve or signal wave.  Extreme values of systolic blood pressure (SBP) and diastolic blood pressure (DBP) can be found on a raw-data daily blood pressure chart.  A “blowup” diagram for a specific annual period of time, e.g. from early 2014 to early 2015, makes it easier to observe certain special circumstances which caused abnormal high BP responses.  A 90-days moving average for both glucose and BP can function as a filter to smooth out certain “noises” (e.g., isolated spikes); therefore, it would be more obvious to observe the “trend” of the curve.  The filtered moving average curves can effortlessly identify the important correlation coefficient between two signal waves of glucose and BP using least square mean algorithm.

Occasionally, a spatial analysis (without “time” factor) is necessary and conveniently used to disclose a hidden “quasi-linear” equation between these two variables.  During 1854, Dr. John Snow created a map-based spatial analysis technique for his Broad Street Cholera Breakout Study in London (Figure 5).  The author got an idea after reading the medical history about Dr. Snow’ work performed 164 years ago and decided to apply a similar concept in his glucose research.  In addition, he has included modern tools from computer science, numerical analysis, big data analytics, and artificial intelligence (AI) in his study.

Furthermore, in order to observe and distinguish certain relationship and trend about glucose and BP, the author selected a shorter period of 750 days (2/8/2014 – 2/27/2016) and a longer period of 1,603 days (2/8/2014 – 6/30/2018) to conduct his detailed comparison analysis.  By comparing the vital information in these two periods such as differences on variables’ correlation and response delay, density and angle of inclination of “data cloud”, as well as the linear governing equation’s surrounding data variance, the information can be observed clearly on the charts of two different time lengths.

In this paper, a total of 9,618 data of six “high-level elements (i.e. FPG, PPG, Daily Glucose, SBP, DBP, M3) based on a period of 1,603 days (2/8/2014 – 6/30/2018) were used for analysis, not including second-tier elements such as food, salt, exercise, weather, etc.  As a matter of fact, several second-tier element analyses already have been performed by the author and will be discussed in a different paper.

Shorter Period (Figure 1) and Longer Period (Figure 2) display the results of BP using both time-series and spatial analyses of glucose vs. BP.  At a quick glance, these two figures appear quite similar.  However, a closer examination will find the following significant differences between the shorter period and longer period:

Figure 1: BP and Glucose of Shorter Period (2/8/2014 - 2/27/2016)
Figure 2: BP and Glucose of Longer Period (2/8/2014 - 6/30/2018)

(1) Correlations between SBP and DBP are 79% for shorter period and 88% for longer period.  This means that the longer period’s results provide a higher level of confidence.  

(2) In general, the glucose and BP curves match with each other quite well in the time-domain of the longer period (2/28/2016 – 6/30/2018).  However, a closer examination can detect that there is a “phase shift” (i.e. time delay) prior to April 2015 and a “near-perfect match” existed after April 2015.  This is due to the patient struggling with his lifestyle change to control his diabetes conditions prior to 4/2015.  This “turbulent period” can be identified easily in Figure 4.  The patient successfully implemented his stringent diabetes control program during the period after 4/11/2015 (release date of weight prediction model) and 6/1/2015 (release date of PPG prediction model).  As shown in Figure 4, it is clear that the BP’s reaction is lagging behind glucose, i.e. with a slower responding speed during this “turbulent” period.  However, starting from April 2016, after following a strict lifestyle management for a year, the BP’s trend started to follow the glucose’s pattern exactly like its twin.

(3) The +/- 20% data coverage area from spatial analysis of glucose and BP are 98.5% for the shorter period and 99.6% for the longer period.  The “denser data cloud” means a “higher data concentration” which provides a clear view of the accuracy of both this linear governing equation and its data concentration band to describe the relationship between glucose and BP.  The data concentration band with two widths of +/- 10% and +/- 20% provides the statistical coverage of data variance and suitability.  This is the reason the author calls this equation a “quasi-linear” equation between glucose and BP.

(4) As shown in Figure 3, the patient had many SBP and DBP spikes in 2014 due to two connected stressful periods caused by family members’ illnesses. After the spring of 2015, his BP (M3) has been stabilized into a healthy state (M3<1.0).  In the longer period diagram, a higher correlation of 88% existed between SBP and DBP with an average of 42 mmHG gap in between.  The time-series analysis of 90-days moving average of glucose vs. BP (M3) also showed a strong correlation of 84%.  This strong correlation was further validated by a spatial analysis which showed 87.7% and 99.6% of the total collected data covered by these two (+/- 10% and +/- 20%) data concentration or data variance bands respectively.

Figure 3: Blood Pressure in 2014 (higher BP Period)
Figure 4: BP lagging behind Glucose (prior to 4/2015)

(5) As shown in Figure 5 and Figure 6, the “straight-line” of governing equation stretched from point A (90, 0.8) to point B (190, 1.06) on a diagram with coordinates of x = glucose, and y = BP (M3).

Figure 5: Spatial Analysis of Glucose vs. BP (no time factor)
Figure 6: Combination of two adjacent periods to form a dense data cloud
which provides a clear view of relationship between Glucose and BP

By using big data analytics on an overweight patient (BMI between 24.7 to 31.0), results show a strong relationship existing between glucose and blood pressure.  This phenomenon has provided a partial answer to one of the author’s question:  “Why did his hypertension go away after focusing and succeeding with his diabetes control?”  This study proves the possible connectivity between diabetes and hypertension.

The author received an honorable PhD in mathematics and majored in engineering at MIT.  He attended different universities over 17 years and studied 7 academic disciplines.

He has spent 20,000 hours in T2D research.  First, he studied six metabolic diseases and food nutrition during 2010-2013, then conducted his own diabetes research during 2014-2018.  His approach is “quantitative medicine” based on mathematics, physics, optical and electronics physics, engineering modeling, signal processing, computer science, big data analytics, statistics, machine learning, and AI.  His main focus is on preventive medicine using prediction tools.  He believes that the better the prediction, the more control you have.


  1. Hsu, Gerald C. (2018). Using Math-Physical Medicine to Control T2D via Metabolism Monitoring and Glucose Predictions. Journal of Endocrinology and Diabetes, 2018(1), 1-6. Retrieved from 2018/06/JEAD-101-1.pdf
  2. Hsu, Gerald C. (2018, June). Using Math-Physical Medicine to Analyze Metabolism and Improve Health Conditions. Video presented at the meeting of the 3rd International Conference on Endocrinology and Metabolic Syndrome 2018, Amsterdam, Netherlands.

The author created this “math-physical medicine” approach by himself in order to save his own life.  Although he has read many medical books, journals, articles, and papers, he did not specifically utilize any data or methodology from other medical work.  All of his research is his original work based on data he collected from his body and using his own computer software developed during the past 8-years.  Therefore, no major problems were associated with data interference or data contamination.  In addition, his knowledge, information, technique, and methodology of mathematics, physics, engineering, and computer science came from his lifelong learning from schools and industries and should not be listed as medical references.   This is the reason his references only contain his own published papers.

Limitation of Research
This article is based on data of metabolic conditions and lifestyle details collected from one T2D patient (himself).  It does not cover genetic conditions and lifestyle details of other diabetes patients.  However, the author’s research approach is based on his solid inter-disciplinary academic background and successful multiple industrial experiences.  His academic background and working experience have prepared him to conduct his diabetes research with the following thorough process and carefully chosen  steps:

  • observing and identifying a system’s basic characters like a pure physicist;
  • developing related but rigorous mathematical equations like a mathematician;
  • applying suitable engineering models and useful statistical models to address the real-world challenges;
  • using modern computer science tools and sophisticated AI techniques to aid in problem solving.

Nevertheless, his conclusions and findings should be re-verified and proceed with caution when applying to other patients who are under different metabolic conditions or lifestyles.

Other Declarations
During the past 8 years of self-study and research, the author has never hired any research assistant or research associate to help with his work except for a part-time computer programmer (~3 evening hours per day) to focus on Apple iOS annual upgrades and system interface problems.  He applied his own invention of a “Software Robot” created during 2001-2009 and AI knowledge he learned to produce the architecture and structure of his customized computer software.  He uses this software to collect and analyze his big data, conduct his medical  research, and then control his diabetes disease.

This project was 100% self-funded by using his own money that was earned from a successful high-tech venture in Silicon Valley.  He did not receive any financial assistance or grants from any public or private institution or organization.  Therefore, there are no concerns regarding any conflict of interest.

First and foremost, the author wishes to express his sincere appreciation to a very important person in his life, Professor Norman Jones at MIT and University of Liverpool.  Not only did he give him the opportunity to study for PhD at MIT, but he also trained him extensively on how to solve difficult problems and conduct any basic scientific research with a big vision, clear head, pure heart, integrity, and honesty.

The author would also like to thank Professor James Andrews at the University of Iowa.  He helped and supported him tremendously when he first came to the United States.  He believed in him and prepared him to build his solid engineering and computer science foundation with a combination of warm heart and strong push during his undergraduate and master’s degree work at Iowa.