Mathematical Translation of Metabolic Syndrome: Assessment of siMS Score for Metabolic Syndrome and Biochemical Risks ()
1. Introduction
Since the conception of the term “metabolic syndrome” various nomenclatures have been used with multiple criteria. Overtime these criteria were refined into more systematic criteria by various organizations like National Cholesterol Education Program (NCEP), World Health Organization (WHO) and International Diabetic Federation (IDF) [1] [2] [3]. Furthermore, regional criteria for labeling subjects with metabolic syndrome also appear with slight variations [4]. While being useful in many ways, there is a tangible gap between these criteria and lack of mathematically defined cut-off to unify metabolic syndrome. Recently, Soldatovic et al. have provided a mathematical model using parameters included in metabolic syndrome, which seems to be promising [5]. However, no local or regional study is available in literature to validate usefulness of this model in our set up.
Insulin resistance:
Why there is a need to have a simplified mathematical model in clinical practice? As highlighted above multiple criteria are there which are overlapping with slightly different cut-offs and parameters which create confusion in routine clinical practice, which may not be providing a unified diagnostic system for patients [6]. Underlying insulin resistance by using insulin-based models HOMAIR is also difficult to measure both due to cost and clinical difficulties in routine practice which makes some of the metabolic syndrome ambiguous in terms of correlation with established metabolic syndrome criteria [7]. An unmet and pushing need therefore arises to have a mathematical defined, easy to a measure and apply criteria for clinical practice.
Earlier attempts to mathematical model metabolic syndrome are also available. Gurka et al. have evaluated race or ethnicity specific mathematical model to predict risk in both adult and pediatric population by utilizing a “confirmatory factor analysis” in SAS program [8]. Similarly, Huh et al. have suggested a clinically applicable equation for continuous monitoring of metabolic syndrome risk for Korean population [9].
This study aims to evaluate the siMS scores among our sample population identified to have metabolic syndrome or otherwise as per IDF criteria and also to correlate this scoring with various biochemical risk factors including insulin resistance.
2. Methods
Study settings and design: This cross-sectional study was carried out at the department of medicine and pathology Naval Hospital (Islamabad) in liaison with chemical pathology department at AFIP for analysis of some biochemical parameters. Referrals were also made by neighboring hospital for possible inclusion into study. The study was comparative cross-sectional and was conducted from July-2017 to Jan-2019. The study has the organizational ethical review committee approval and ensured written signed consent for all participants.
Target population & subject selection: All adults visiting hospital for executive annual medical check-up considered as target population. Patient selection was based upon the “non-probability convenience sampling” methodology. Subjects with known metabolic disorder like diabetes, hypertension, ischemic heart disease (IHD), autoimmune disorder or other acute or chronic ailments were excluded from the study. Pregnant subjects were also excluded from the study.
Sampling & general clinical examination: The study participants were requested to visit in medical fasting at the department of pathology around 08:00 to 09:00 hours from Monday to Friday. These patients were explained the medical fasting requirements along with an explanation regarding sampling strategy, purpose of research project and post-results utilization purpose of data. The subjects who finally appeared (N = 232) for study as per conditions explained to them at pathology department were fist asked to sign the written consent form. This was followed by a general history and examination to rule any signs of a chronic disease process. After this the phlebotomist collected up to 10 ml of blood from all study subjects for various biochemical parameters including fasting glucose and lipids, HbA1c and insulin in specified containers. A urine sample was collected from 171 subjects for measuring urine albumin creatinine ratio (UACR).
Clinical measurements and analysis: All anthropometric measurements were calculated as per the WHO defined criteria and Guerrero-Romero et al. methodology for abdominal volume index (AVI) [10] [11]. Biochemical parameters including cholesterol, triglycerides and glucose were measured by CHOD PAP, GPO PAP and GOD PAP techniques on random access clinical chemistry analyzer (Slectra ProM). HDL and LDL lipoproteins were measured using cholesterol esterase methodology on AVIDA-1800 random access clinical chemistry analyzer. HbA1c was analyzed using ion-exchange resin chromatography method. Serum insulin was analyzed using Immulite® 1000 chemiluminesence analyzer.
Outcome measures:
· Metabolic syndrome and diabetes was diagnosed as per IDF criteria, as:
○ Waist circumference ≥ 94 cm (Males) and ≥80 cm (Females), PLUS ANY OF TWO as:
○ HDLc < 1.03 mmol/L (Males) or <1.3 mmol/L (Females);
○ Triglycerides > 1.7 mmol/L;
○ Blood pressure > 130/85 mm of Hg;
○ Fasting glucose 5.6 mmol/L [12] [13].
· siMS score was calculated as per Soldatovic et al. formula [5], as described below:
Male:
siMS score = (2 × Waist/Height) + (Glucose/5.6) + (Triglyceride/1.7) + (TAsystolic/130) − (HDL/1.02).
Female:
siMS score = (2 × Waist/Height) + (Glucose/5.6) + (Triglyceride/1.7) + (TAsystolic/130) − (HDL/1.28).
· HOMA IR was calculated by the formula of Mathew’s et al. as:
HOMA IR = [{Serum Insulin (mIU/L)} × {Fasting Plasma Glucose (mmol/L)}]/22.5 [14].
Quality control & calibration: Internal and external laboratory testing is ensured by both external QC insurance program like “National External Quality Assurance Program Pakistan (NEQAPP)” and internal by regular monitoring and troubleshooting through Westgard’s quality control rules. The usual targets for precision (%CV) and accuracy for both inter and intra batch testing are ensured on regular basis along with documentation of any error.
We lost few samples while processing and could not follow up for serum insulin (N = 4), HbA1c (N = 2), and UACR testing was only done in 174 subjects.
Statistical analysis: All data were added to Excel program and later moved to IBM-SPSS version-24. Age and gender based differences were calculated through descriptive statistics option in SPSS while the gender group wise differences between measured and calculated parameters were carried out through Independent Sample t-statistics. Independent sample t-test was used to see the differences for siMS score and insulin resistance between subjects with and without metabolic syndrome. Bivariate Pearson’s correlation method between various evaluated anthropometric and biochemical risk assessment parameters. ROC curve analysis was used to measure the Area Under Curve (AUC) for various parameters in diagnosis of metabolic syndrome.
3. Results
Out of the total sample size we had 52.4% females and 47.6% males. Gender wise descriptive data is shown in Table 1 for age, anthropometric indices, biochemical measures and siMS score. Insulin resistance between study subjects is as: Metabolic syndrome subjects: Mean = 3.27 ± 4.45 and non-metabolic syndrome subjects: Mean = 2.10 ± 1.89 (p = 0.012) as shown in Figure 1. Figure 2 shows the differences in siMS score to be higher in subjects with metabolic syndrome (Mean = 3.58 ± 0.725, N = 121) than subjects without metabolic syndrome (Mean = 2.83 ± 0.727, N = 108) as per IDF defined criteria. Pearson’s correlation between various biochemical risk predictor and siMS score suggests siMS score to be highly correlated with anthropometric, glycemic, lipid indices and insulin resistance than other parameters (Table 2). AUC for various parameters for various biochemical parameters include sdLDL cholesterol (AUC = 0.700), non-HDL
![]()
Table 1. Gender differences between subjects for various evaluated parameters.
*Independent Sample t-test.
![]()
![]()
Table 2. Correlation between non-HDL cholesterol, small dense LDL cholesterol (sdLDLc), fasting plasma glucose (FPG), homeostasis model assessment for insulin resistance (HOMA IR), urine albumin creatinine ratio (UACR) and siMS score.
*p-value < 0.05. **p-value < 0.01.
![]()
Figure 1. Differences in HOMAIR in subjects with and without metabolic syndrome (p = 0.012).
![]()
Figure 2. Differences in siMS score in subjects with and without metabolic syndrome (p < 0.001).
cholesterol (AUC = 0.647), HbA1c (AUC = 0.644), Fasting plasma glucose (AUC = 0.698), HOMA IR (AUC = 0.629) and siMS score (AUC = 0.866) as depicted in Figure 3.
4. Discussion
siMS score appeared promising in terms of mathematical translation of data from various risk biomarkers included in the definition of metabolic syndrome. Our study also confirmed that this mathematical scoring method is highly correlated with insulin resistance along with demonstrating highest area under the curve for diagnosing metabolic syndrome. Using such an index can help add to diagnostics along with its application as a measurable tool to assess the overall improvement or further worsening of the disease in a patient’s prognostic monitoring. While being a simple mathematical number, it can provide the combined data from components of metabolic syndrome to simply patient management. However, Soldatovic et al.’s as we demonstrated seems to be very useful and in this regard there are few studies which support its use in clinical care [5]. Vukovic et al. have evaluated this method in pediatric population with slight modifications to conclude it as an accurate and practical measure in diagnosis of metabolic syndrome in children and adolescence [15]. Similarly, another pediatric study “the CASPIAN-V study” have also highlighted the practical usefulness in both clinics and research programs after comparing it with various principal component analysis, confirmatory component analysis and z-scores [16].
Provided clinical utility some researchers have highlighted some issues which may not be specific to siMS score, but still needs to be highlighted while using siMS score as a biomarker for metabolic syndrome. Srećković et al. have shown
![]()
Figure 3. Receiver operating curve (ROC) curve analysis showing area under curve (AUC) for siMS score, HOMAIR, Fasting plasma glucose, HbA1c, non-HDL cholesterol and small dense LDL cholesterol (sdLDLc).
that certain acute phase reactants and cardiovascular disease (CVD) indicators like C-Reactive protein (CRP), Fibrinogen, acidum uricum, Apo-B lipoprotein and homocysteine can confound the siMS score risk prediction [17]. Furthermore Sebekova et al. have shown that real-time application of siMS score may label some of the subjects without criterion defined metabolic syndrome to have numerical results falling in the range of metabolic syndrome, thus can result in some false positive cases [18]. However, in the opinion of authors we agree with these later findings as metabolic risk clustering with emerging evidence is ever evolving field and more and newer CVD biomarkers depictive of underlying inflammation and atherosclerotic plaque behavior are now entering clinical market [19]. Moreover, the role of genetics and epigenetics in the absence or presence of metabolic syndrome label, as per the newer evidence is probably more informative in risk prediction than simply quantifying biochemical risk predictor [20] [21].
The study has few limitations. The authors attempted to replicate the use of siMS score in our population; however our sample size was small and the study was hospital based so a larger study in an epidemiological set up may be carried out for validating the siMS score equation for primary care set up. Moreover, as per the findings of Sebekova et al. we feel metabolic syndrome could have multifactorial etiology beyond what we have included in our study which thus raises the need of other factors which could lead to false negative low siMS score in the presence of a metabolic syndrome. Clinical assessment therefore must be personalized and this consideration must remain in the mind of treating physicians.
The study has significant clinical implications: The authors consider the use of siMS score evaluation could provide a real-time assessment model for primary care physician who can not only use this equation as “Rule In” investigation for metabolic syndrome, which can also be used to monitor the treatment and assessment of prognosis for patients. Furthermore the mathematical equation may also be incorporated in hospital information system so as to calculate by the IT system to allow its simplistic application for both patients and physicians.
5. Conclusion
siMS score has shown better performance than HOMAIR, sdLDL cholesterol, non-HDL cholesterol, HbA1c, and fasting plasma glucose in diagnosing metabolic syndrome. Furthermore, the siMS score equation has moderate to high correlation with most of the anthropometric, biochemical and hormonal risk factors.
Declarations
Ethical approval: The study “Mathematical translation of metabolic syndrome by using siMS score: Assessment of siMS score for metabolic syndrome and biochemical risks” was formally approved by hospital’s ethical review committee.
We confirm that our research work conforms to “World Medical Association’s Declaration of Helsinki—Ethical Principles for Medical Research Involving Human Subjects”
Signing of inform consent by participants: All subjects were required to sign the written informed consent before inclusion into study program. All included subjects were explained about the research requirements and use of data along with confidentiality issues.
Availability of SPSS data & outputs: Subject data can be provided on formal request.
Author’s Contributions
SHK: (Author for correspondence) Study idea, study conceptualization, Diagnostic lab analysis, Manuscript writing, Study finalization. ARK: Study sampling, analysis of data, manuscript writing. AH: Patient selection, examination, referrals, data output (SPSS) analysis, data analysis and manuscript writing. RA: Clinical evaluation of patient, statistical methods application & analysis, manuscript writing. RS: Sampling collection, data analysis, contribution to manuscript write up. TC: Overall study coordination, medical writing, study finalization. All study authors approved the final manuscript version and agreed to all aspects of contents.
Data Funding
There is no funding source to declare.
Acknowledgements
The authors want to acknowledge assistance provided by Lab technician Ibrahim and Iftikhar.
Abbreviations
Homeostasis Model Assessment for Insulin Resistance (HOMA-IR), siMS Score, Small Dense Low Density Lipoprotein Cholesterol (sdLDLc), Urine Albumin Creatinine Ratio (UACR), Body Mass Index (BMI), Waist to Hip Ratio (WHpR), Waist to Height Ratio (WHtR), Abdominal Volume Index (AVI)