• Users Online: 78
  • Print this page
  • Email this page


 
 
Table of Contents
ORIGINAL ARTICLE
Year : 2019  |  Volume : 9  |  Issue : 1  |  Page : 7-16

Statistical correlation of severity of coronary artery disease with insulin resistance and other clinical parameters


1 Department of Biomedical Sciences and Engineering, Tampere University of Technology, Tampere, Finland
2 Department of Electrical and Electronics Engineering, Manipal Institute of Technology, Manipal, Karnataka, India
3 Department of Cardiology, Kasturba Medical College, Manipal, Karnataka, India

Date of Web Publication10-May-2019

Correspondence Address:
Mr. Jay Rabindra Kumar Samal
Department of Biomedical Sciences and Engineering, Tampere University of Technology, P. O. Box: 527, FI-33101, Tampere
Finland
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/JICC.JICC_1_18

Rights and Permissions
  Abstract 


Objective: The objective of the study is to find the correlation of severity of coronary artery disease (Gensini score) with insulin resistance (IR) and other clinical parameters. Methods: Clinical data set (Gensini score, glycated hemoglobin, fasting insulin (FI), IR, total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein, age, body mass index, waist circumference, high-sensitivity c-reactive protein and fasting plasma glucose) of 100 patients was collected. The individuals included in the data set were classified into four groups based on IR and phenotypic obesity. R programing language was used to find correlation between the clinical parameters and the Gensini score. Further, the data were normalized and data plot on cftools of Matlab was used to find equations to relate the parameters. The variation of Gensini score among the four groups was also analyzed. Results: The variation of Gensini score among the four groups suggests that IR can drastically increase the severity of CAD and has a more pronounced effect on the Gensini score as compared to phenotypic obesity. It was also observed that age, triglyceride levels, glycated hemoglobin, and FI had the highest positive correlation with the Gensini score, while parameters such as body mass index and high-sensitivity C-reactive protein had a higher negative correlation. Equations correlating various clinical parameters to the Gensini score were generated. Conclusion: The equations may be used to develop a software model which can predict the Gensini Score of a patient with an acceptable margin of error. As the parameters included in the study can be obtained during a regular check-up of a person, his/her risk of CAD can be predicted during the regular check-up.

Keywords: Coronary artery disease, Gensini score, insulin resistance


How to cite this article:
Samal JR, Sinha A, Uppunda D. Statistical correlation of severity of coronary artery disease with insulin resistance and other clinical parameters. J Indian coll cardiol 2019;9:7-16

How to cite this URL:
Samal JR, Sinha A, Uppunda D. Statistical correlation of severity of coronary artery disease with insulin resistance and other clinical parameters. J Indian coll cardiol [serial online] 2019 [cited 2019 Jul 22];9:7-16. Available from: http://www.joicc.org/text.asp?2019/9/1/7/257951




  Introduction Top


Coronary artery disease (CAD) is caused due to the build-up of plaque along the arteries of the heart, which results in narrowing of the arteries and subsequent reduction in blood flow to the heart.[1] The Gensini score evaluates the number of stenotic artery segments and the degree of lumen stenosis and is considered the standard score to signify the severity of CAD in a patient.[1],[2],[3]

The South Asian region is home to more than 20% of the world's population.[4] Insulin resistance (IR) syndrome is a condition highly prevalent and oft-overlooked in the Indian subcontinent and even affects adolescents.[5],[6] Various studies have suggested the possibility of an Indian phenotype, which predisposes Indians to the development of IR.[7] It has also been suggested that IR can significantly increase the risk of CAD.[7] IR has been given greater importance in this study because albeit the majority of the Indian population is suffering from IR, it is often ignored as physicians rather concentrate on phenotypic obesity to find the associated risk of diabetes.

This study aimed at finding correlations between the Gensini score and clinical parameters. Equations representing the relationship between the Gensini score and clinical parameters were developed. A software model can be developed using these equations, which will evaluate the Gensini score of a patient by using clinical parameters that can be obtained during a regular check-up.


  Methods Top


Clinical data set (Gensini score [severity of CAD], glycated hemoglobin [HBA1C], fasting insulin [FI], homeostatic model assessment IR [HOMA_IR], total cholesterol [TC], triglycerides [TG], high-density lipoprotein [HDL], low-density lipoprotein [LDL], age, body mass index [BMI], waist circumference [WC], high-sensitivity C-reactive protein (hs-CRP), and fasting plasma glucose 7) of 100 patients was collected from Kasturba Medical College, Manipal, India. The study was approved by the Institutional Ethics Committee of Kasturba Medical College, Manipal, India. The patients showed general symptoms of CAD such as having chest pain and dyspnea on exertion. None of the patients were under statin treatment.

The individuals included in the data set were classified into four groups based on IR and phenotypic obesity as follows:

  • Group 1: Metabolically healthy normal weight (IR negative, obesity negative)
  • Group 2: Metabolically obese normal weight (MONW) (IR positive, obesity negative)
  • Group 3: Metabolically healthy obese (IR negative, obesity positive)
  • Group 4: Metabolically abnormal obese (IR positive, obesity positive).


R programing language was used to find a correlation between the clinical parameters and the Gensini score. Correlation tables were created for each of the four groups, which give an idea regarding the relative increase or decrease of the Gensini score with different clinical parameters.

Further, the data were normalized, and data plot on cftools of Matlab was used to find equations to relate the parameters by manually verifying the best possible fit of plethora of combinations possible. In case, an accurate equation could not be obtained, the outlying data were removed, and the best possible fit was verified again.

The variation of Gensini score among the four groups was also analyzed.


  Results Top


The Gensini scores were highest for individuals in Group 4 (IR positive and obesity positive). Surprisingly, individuals included in Group 3 (IR negative and obesity positive) had the lowest Gensini scores among the four groups while individuals of Group 2 (IR positive and obesity negative), despite being phenotypically normal, had higher Gensini scores. This signifies that IR can drastically increase the severity of CAD and has a more pronounced effect on the severity of CAD as compared to phenotypic obesity [Chart 1].



The numbers represent the extent of positive/negative effect of each clinical parameter on the Gensini score on a scale from −1 to +1. These correlations were obtained by using the R programing language [Table 1].
Table 1: Correlation between clinical parameters and Gensini score

Click here to view


Equations correlating various clinical parameters to the Gensini score were further generated using data plot on cftools of Matlab.

y = p1X6 + p2X5 + p3X4 + p4X3 + p5X2 + p6x + p7R2 = 1(1)

p1= −4.237e + 004 p2= −1574 p3 = 1.484e + 004 p4 = 1525

p5= −953.1 p6= −48.01 p7 = 15.98

Equation 1 shows sample equation correlating BMI to Gensini Score of Group 2 (MONW).

The accuracy of the equations is indicated by the R2 value, which indicates goodness of fit. The R2 value of almost all the equations generated was found to be 1, which indicates that the clinical parameters were forming a perfect curve with the Gensini score and accurately represent the relationship between the parameters and the Gensini score. The complete list of equations generated (for each parameter and each group) has been attached as supplementary material.


  Discussion Top


The variation of Gensini score among the four groups suggests that IR can drastically increase the severity of CAD and has a more pronounced effect on the Gensini score as compared to phenotypic obesity.

Parameters such as Age, HbA1c, and FPG had the highest positive correlation, while parameters such as BMI and FI showed the highest negative correlation with the Gensini score for individuals in Group 1.

For Group 2, TG, Age, HbA1c, FI, and WC had the highest positive correlation, while BMI, HDL, LDL, and hsCRP showed the highest negative correlation with the Gensini score.

Age, FPG, FI, HOMA_IR, and TG showed the highest positive correlation with Gensini score in Group 3. On the other hand, BMI and hsCRP had a higher negative correlation.

Age, FPG, FI, HOMA_IR, and hbA1c showed a higher positive correlation with Gensini score in Group 4 while TG, WC, and TC showed the highest negative correlation [Chart 2].




  Conclusion Top


A coronary angiogram is currently the only method available to measure the severity of CAD in a patient. An angiogram is quite expensive, often not affordable and may also lead to various complications.[8] Moreover, many patients are reluctant to undergo an angiogram even on the doctor's insistence as they are scared of the complications that may arise. Most often, CAD is not detected until the patient has already suffered a heart attack as angiograms are not opted for in the case of seemingly normal controls, due to which most of the cases go undiagnosed and are oft fatal.

Once the equations have been generated, it may be possible to develop a software model, which is beyond the scope of this paper and may predict the Gensini Score of a patient with an acceptable margin of error. This model may be continuously improved by adding more sample data and taking into consideration a large community or clinical data set so as to generate more accurate equations.

As the parameters included in the study can be obtained during a regular check-up of a person, his/her risk of CAD can be predicted during the regular check-up so that the person can be treated before he develops further complications, which can be life-threatening in most instances.

Acknowledgment

We would like to thank Philips India Limited for providing us with the dataset used for our analysis.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Group 1: Metabolically Healthy Normal Weight (MHNW) (IR negative, Obesity negative)

1. Low Density Lipoprotein (LDL) R2=0.831

y=a1sin (b1x+c1) +a2sin (b2x+c2) +a3sin (b3x+c3)+a4sin(b4x+c4)+a5sin(b5x+c5)+a6sin (b6x+c6) +a7sin (b7x+c7)

a1=5.882 b1=91.01 c1=-0.08961

a2=6.807 b2=43.11 c2=0.2057

a3=10.94 b3=66.11 c3=5.561

a4=10.94 b4=9.916 c4=0.3681

a5=9.66 b5=52.95 c5=2.694

a6=9.705 b6=82.11 c6=2.826

a7=10.79 b7=26.21 c7=3.613

2. Insulin Resistance (Homa_IR) R2=0.8302

y=a1sin(b1x+c1)+a2sin(b2x+c2)+a3sin(b3x+c3)+a4sin(b4x+c4)+a5sin(b5x+c5)

+a6sin(b6x+c6)

a1=2.832 b1=134.8 c1=-0.4419

a2=2.492 b2=166.1 c2=1.092

a3=1.241 b3=199.7 c3=-2.084

a4=3.962 b4=102.5 c4=-2.523

a5=2.19 b5=36.71 c5=-0.3184

a6=3.309 b6=65.99 c6=1.04

3.Fasting Plasma Glucose (FPG) R2=0.8103

y=a1sin(b1x+c1)+a2sin(b2x+c2)+a3sin(b3x+c3)+a4sin(b4x+c4)+a5sin(b5x+c5)

+a6sin(b6x+c6)+a7sin(b7x+c7)+a8sin(b8x+c8)

a1=46.33 b1=5.423 c1=-2.204

a2=90.72 b2=18.25 c2=-0.5102

a3=108.3 b3=25.12 c3=-5.637

a4=30.03 b4=43.25 c4=-1.668

a5=96.2 b5=31.06 c5=1.661

a6=6.526 b6=51.2 c6=3.034

a7=70.69 b7=12.51 c7=1.96

a8=59.79 b8=37.29 c8=0.03115

4.Glycated Hemoglobin (HbA1c) R2=0.8596

y=a1sin(b1x+c1)+a2sin(b2x+c2)+a3sin(b3x+c3)+a4sin(b4x+c4)+a5sin(b5x+c5)

+a6sin(b6x+c6)+a7sin(b7x+c7)+a8sin(b8x+c8)

a1=2.494 b1=6.283 c1=0.7999

a2=0.4687 b2=37.72 c2=-0.03522

a3=1.433 b3=18.87 c3=-0.8301

a4=1.589 b4=3.123 c4=-1.579

a5=0.1625 b5=50.24 c5=2.815

a6=0.9919 b6=31.41 c6=-2.853

a7=2.091 b7=12.57 c7=3.038

a8=0.9906 b8=25.13 c8=1.091

Group 2: Metabolically Obese Normal Weight (MONW) (IR positive, Obesity negative)

1. Body Mass Index (BMI) R2=1

y= p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = -4.237e+004 p5 = -953.1

p2 = -1574 p6 = -48.01

p3 = 1.484e+004 p7 = 15.98

p4 = 1525

2. Waist Circumference (WC) R2=0.9981

y= p1x5 + p2x4 + p3x3 + p4x2 + p5x + p6

p1 = 2314 p4 =-2.552

p2 = -409.9 p5 = 26.95

p3 = -688.1 p6 = -0.5024

3. Glycated Hemoglobin (HbA1c) R2=1

y= p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = -1090 p4 = 63

p2 = -48.39 p5 = -46.24

p3 = 495.7 p6 = -2.41

p7 = 1.291

4. Fasting Plasma Glucose (FPG) R2=1

y= p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = -1.879e+004 p4 = -302.9

p2 = 6830 p5 =-397.6

p3 =5854 p6 =-34.16

p7 = 0.2192

5. Triglycerides (TG) R2=1

y= p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = -6210 p4 = 155.4

p2 =-2075 p5 = -117.3

p3 = 1932 p6 = -2.74

p7 = 1.444

6. High Density Lipoprotein (HDL) R2=1

y=p1x5 + p2x4 + p3x3 + p4x2 + p5x + p6

p1 = 560.3 p4 = 50.15

p2 = -293 p5 =10.14

p3 = -141.7 p6 = -0.2155

7. Low Density Lipoprotein (LDL) R2=1

y= p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 =1111 p4 = -183.5

p2 =653.3 p5 = 19.82

p3 = -332.8 p6 = 7.556

p7 = -0.5091

8. CH-HDL R2=1

y=p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = -1425 p4 = 212.5

p2 =-1150 p5 = 4.45

p3 = 141.4 p6 =-6.326

p7 = -0.119

9. Fasting Insulin (FI) R2=1

y=p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = 1059 p4 = 45.93

p2 =-203 p5 = 34.17

p3 =-370.4 p6 = -1.155

p7 = -1.08

10.Insulin Resistance (HOMA_IR) R2=1

y=p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = 4.045e+004 p4 = -2386

p2 =-1539 p5 =997.3

p3 = -1.535e+004 p6 = 260

p7 =13.78

11.High sensitive CRP (hsCRP) R2=1

y=p1x6 + p2x5 + p3x4 + p4x3 + p5x2 + p6x + p7

p1 = -1.119e+005 p4 = -8314

p2 = 7.337e+004 p5 =-307.5

p3 = 1.752e+004 p6 =255.8

p7 = -16.5

Group 3: Metabolically Healthy Obese (MHO) (IR negative, Obesity positive)

1. Body Mass Index (BMI) R2=9.997

y= p1x5+p2x4+p3x3+p4x2+p5x+p6, where

p1=2419 p4=-47.14

p2=774.3 p5=1.256

p3=-182.3 p6=0.0719

2. Waist Circumference (WC)

y=p1x2+p2x+p3

p1=102.4 p2=-8.747 p3=-0.1182

3. Fasting Plasma Glucose (FPG) R2=1

y=p1x5+p2x4+p3x3+p4x2+p5x+p6

p1=1670 p4=183.8

p2=-913.2 p5=19.57

p3=-344.7 p6=-8.691

4. Total Cholesterol (TC) R2=0.997

y=p1x3+p2x2+p3x+p4

p1=-2289 p3=-88.7

p2=908.4 p4=1.079

5. Triglycerides (TG) R2=0.997

y=a0+a1cos(xw)+b1sin(xw)+a2cos(2xw)+b2sin(2xw)

a0=-2.881e+0y b1=-8,649e+0y

a1=0 b2=4.271e+0y

a2=2.88e+0y w=-0.05318

6. High Density Lipoprotein (HDL) R2=0.992

y=a0+a1cos(wx)+b1sin(wx)+a2cos(2wx)+b2sin(2wx)

a0=2.345 b1=-3.217

a1=3.016 b2=-2.469

a2=-0.9426 w=7.556

7. Low Density Lipoprotein (LDL) R2=0.8735

y=a0+a2cos(3xw)+b3sin(3xw)

a0=0.3152 b3=-0.8629

a3=-0.001229 w=17.8

8. Fasting Insulin (FI) R2=0.9973

y=p1x6+p2x5+p3x4+p4x3+p5x2+p6x+p7

p1=-5.287e+04 p5=-1696

p2=2948 p6=-252.7

p3=1.984e+04 p7=-5.717

p4=692

9. Insulin Resistance (Homa_IR)

y=p1x6+p2x5+p3x4+p4x3+p5x2+p6x+p7

p1=2539 p4=-76.98

p2=35.87 p5=94.45

p3=-1005 p6=15.78

p7=-0.005868

Group 4: Metabolically Abnormal Obese (MAO) (IR positive, Obesity positive)

1. Glycated Hemoglobin (HbA1c) R2=1

y=a1sin(b1x+c1) + a2sin(b2x+c2) + a3sin(b3x+c3) + a4sin(b4x+c4)

a1 = 0.5058 b1 = 20.6 c1 =0.3266

a2 =0.7804 b2 = 23.27 c2 = -2.188

a3 = 0.5373 b3 = 3.367 c3 = -1.531

a4 =0.4573 b4 = 32.63 c4 = -2.118

2. Fasting Plasma Glucose (FPG) R2=0.9965

y=p1x9 + p2x8 + p3x7 + p4x6 + p5x5 + p6x4 + p7x3 + p8x2 + p9x + p10

p1 = -5.647e+004 p6 =1854

p2 = 2.818e+004 p7 = 641.2

p3 = 3.232e+004 p8 = -83.74

p4 = -1.297e+004 p9 = -15.67

p5 =-7005 p10= 1.351

3. Total Cholesterol (TC) R2= 1

y=a0 + a1cos(xw) + b1sin(xw) + a2cos(2xw) + b2sin(2xw) + a3cos(3xw) + b3sin(3xw) + a4cos(4xw) + b4sin(4xw) + a5cos(5xw) + b5sin(5xw) + a6cos(6xw) + b6sin(6xw)

a0 = 2.858e+006 b1 = 0

a1 = 0 b2 = -1.759e+006

a2 = -9.569e+006 b3 = 3.223e+006

a3 = 1.194e+007 b4 = -2.52e+006

a4 = -7.221e+006 b5 = 9.66e+005

a5 = 2.307e+006 b6 = -1.5e+005

a6 = -3.146e+005 w = 1.571

4. Triglycerides (TG) R2=1

y= a1sin(b1x+c1) + a2sin(b2x+c2) + a3sin(b3x+c3) +

a4sin(b4x+c4) + a5sin(b5x+c5) + a6sin(b6x+c6)

a1 = 0.3502 b1 = 6.283 c1 = -0.2775

a2 =0.7256 b2 =12.57 c2 =-0.8475

a3 =0.406 b3 = 3.142 c3 =1.993

a4 = 0.3601 b4 =31.42 c4 = 1.611

a5 =0.4333 b5 =18.85 c5 =-2.41

a6 =0.5931 b6 = 25.13 c6 =2.585

5. Low Density Lipoprotein (LDL) R2=1

y= a1sin(b1x+c1) + a2sin(b2x+c2) + a3sin(b3x+c3) +

a4sin(b4x+c4) + a5sin(b5x+c5) + a6sin(b6x+c6)

a1 = 421.6 b1 = 6.283 c1 =1.466

a2 = 123.1 b2 =18.85 c2 = 2.317

a3 = 48.88 b3 =25.13 c3 =-0.4761

a4 =324.8 b4 =3.142 c4 =-2.125

a5 =11.88 b5 = 31.42 c5 = 2.952

a6 = 230.1 b6 = 12.57 c6 =-1.185

6. Insulin Resistance (HOMA_IR) R2=1

y=a1sin(b1x+c1) + a2sin(b2x+c2) + a3sin(b3x+c3) +

a4sin(b4x+c4) + a5sin(b5x+c5) + a6sin(b6x+c6)

a1 = 7.361e+004 b1 = 25.13 c1 =-1.92

a2 = 1.001e+004 b2 = 31.42 c2 =2.036

a3 = 2.633e+005 b3 = 18.85 c3 =0.3968

a4 = 6.264e+005 b4 =12.57 c4 =2.674

a5 = 1.438e+006 b5 =6.283 c5 =-1.493

a6 = 1.37e+006 b6 = 3.142 c6 =0.944

7. High Sensitive CRP (hsCRP) R2=0.9897

y=a1sin(b1x+c1) + a2sin(b2x+c2) + a3sin(b3x+c3) +

a4sin(b4x+c4) + a5sin(b5x+c5) + a6sin(b6x+c6) +

a7sin(b7x+c7)

a1 = 2.838e+004 b1 =31.49 c1 =2.239

a2 = 2.715e+004 b2 = 12.47 c2 =1.85

a3 = 1.678e+005 b3 = 18.89 c3 =0.02677

a4 = 1.073e+005 b4 =6.314 c4 = 1.092

a5 = 1.213e+005 b5 = 25.2 c5 = -2.01

a6 = 0.01216 b6 =6.283 c6 = 2.336e-006

a7 = 0.01216 b7 = 6.283 c7 = 2.336e-006

8. Fasting Insulin R2=1

y=a1sin(b1x+c1) + a2sin(b2x+c2) + a3sin(b3x+c3) + a4sin(b4x+c4) + a5sin(b5x+c5) + a6sin(b6x+c6)

a1 = 2.212e+005 b1 =18.85 c1 =0.4625

a2 =1.05e+004 b2 =31.42 c2 = 2.372

a3 = 4.526e+005 b3 = 12.57 c3 = 2.677

a4 = 7.001e+004 b4 =25.13 c4 =-1.735

a5 = 8.728e+005 b5 =6.283 c5 = -1.462

a6 = 7.446e+005 b6 = 3.142 c6 = 1.004



 
  References Top

1.
He LY, Zhao JF, Han JL, Shen SS, Chen XJ. Correlation between serum free fatty acids levels and Gensini score in elderly patients with coronary heart disease. J Geriatr Cardiol 2014;11:57-62.  Back to cited text no. 1
    
2.
Neeland IJ, Patel RS, Eshtehardi P, Dhawan S, McDaniel MC, Rab ST, et al. Coronary angiographic scoring systems: An evaluation of their equivalence and validity. Am Heart J 2012;164:547-52.e1.  Back to cited text no. 2
    
3.
Gensini GG. A more meaningful scoring system for determining the severity of coronary heart disease. Am J Cardiol 1983;51:606.  Back to cited text no. 3
    
4.
Ogurtsova K, da Rocha Fernandes JD, Huang Y, Linnenkamp U, Guariguata L, Cho NH, et al. IDF diabetes atlas: Global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes Res Clin Pract 2017;128:40-50.  Back to cited text no. 4
    
5.
Sharp PS, Mohan V, Levy JC, Mather HM, Kohner EM. Insulin resistance in patients of Asian Indian and European origin with non-insulin dependent diabetes. Horm Metab Res 1987;19:84-5.  Back to cited text no. 5
    
6.
Misra A, Vikram NK, Arya S, Pandey RM, Dhingra V, Chatterjee A, et al. High prevalence of insulin resistance in postpubertal Asian Indian children is associated with adverse truncal body fat patterning, abdominal adiposity and excess body fat. Int J Obes Relat Metab Disord 2004;28:1217-26.  Back to cited text no. 6
    
7.
Unnikrishnan R, Anjana RM, Mohan V. Diabetes in South Asians: Is the phenotype different? Diabetes 2014;63:53-5.  Back to cited text no. 7
    
8.
Harris JM Jr. Coronary angiography and its complications. The search for risk factors. Arch Intern Med 1984;144:337-41.  Back to cited text no. 8
    



 
 
    Tables

  [Table 1]



 

Top
 
  Search
 
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

 
  In this article
Abstract
Introduction
Methods
Results
Discussion
Conclusion
References
Article Tables

 Article Access Statistics
    Viewed287    
    Printed31    
    Emailed0    
    PDF Downloaded44    
    Comments [Add]    

Recommend this journal