Comparison of scores for the classification of cardiometabolic risk in adult patients enrolled in a Venezuelan program for chronic non-communicable diseases: a cross-sectional study

ABSTRACT BACKGROUND: Several continuous measurements of cardiometabolic risk (CMR) have emerged as indexes or scores. To our knowledge, there are no published data on its application and validation in Latin America. OBJECTIVE: To evaluate four continuous measurements of metabolic status and CMR. We established its predictive capacity for four conditions associated with CMR. DESIGN AND SETTING: Cross-sectional study conducted at a healthcare center in the state of Carabobo, Venezuela. METHODS: The sample comprised 176 Venezuelan adults enrolled in a chronic disease care program. Four CMR scores were calculated: metabolic syndrome (MetS) Z-score; cardiometabolic index (ICMet); simple method for quantifying MetS (siMS) score; and siMS risk score. CMR biomarkers, proinflammatory status and glomerular function were assessed. MetS was established in accordance with a harmonized definition. RESULTS: Patients with MetS showed higher levels of all scores. All scores increased as the number of MetS components rose. The scores showed significant correlations with most CMR biomarkers, inflammation and glomerular function after adjusting for age and sex. In the entire sample, MetS Z-score, siMS score and siMS risk score showed the ability to detect MetS, reduced glycemic control, proinflammatory status and decreased estimated glomerular filtration rate. ICMet only discriminated MetS and proinflammatory state. There were some differences in the predictive capacity of the scores according to sex. CONCLUSIONS: The findings support the use of the scores assessed here. Follow-up studies should evaluate the predictive capacity of scores for cardiovascular events and diabetes in the Venezuelan population.

factors accumulated. 7,8 Also, the conventional diagnosis of metabolic syndrome does not make it possible to follow up the gradual changes that occur in individuals with metabolic syndrome once the therapeutic measures are in place. 8 A continuous cardiometabolic risk index responds to the above limitations. It shows the continuous risk to which an individual is exposed and provides information about the severity of the risk.
Over recent years, several continuous measurements of cardiometabolic risk have emerged as indexes or scores. In general, these include the same individual components of metabolic syndrome but differ in the methodologies that are used for their construction and calculation.
None of these proposed continuous scores for metabolic syndrome originated from the Latin American population. To our knowledge, there are no published data on application and validation of such scores in Latin America. It is important to consider that the prevalence of this disease, its survival rates and the distribution of risk factors and their weights as determinants of the disease may be different in each population. 9 There is also genetic and environmental control over the expression of cardiometabolic risk factors in each population group.

OBJECTIVE
The aim of this research was to evaluate four continuous measurements of metabolic status and cardiometabolic risk in a group of adult patients who had been enrolled in the CAREMT (Cardio Renal Endocrine Metabolic and Tobacco) program, which was developed at a healthcare center in the state of Carabobo, Venezuela. We explored the variation of continuous measurements according to different biomarkers for cardiometabolic risk, inflammation and glomerular function. We established the ability of continuous measurements to discriminate or detect metabolic syndrome, reduced metabolic control, proinflammatory status and decreased glomerular function. This exploratory assessment was the first step towards validation of continuous measurements of cardiometabolic risk in Latin American countries such as Venezuela, for future primary care applications.

Participants and data collection procedure
This was a cross-sectional study of correlational type, with a non-experimental design. The validation of continuous measurements was performed using a cross-section of baseline data from the CAREMT program, implemented at a primary healthcare center in the state of Carabobo, Venezuela. This program consists of an integration of the cardiovascular, endocrine, metabolic, renal, cancer and anti-smoking programs, in a strategy for screening and prevention of the most frequent non-communicable chronic diseases and their risk factors. The study was based on non-probabilistic sampling. The population comprised all the adult patients (20-65 years of age) of both genders who were enrolled in the CAREMT program between 2015 and August 2017 (n = 210). The sample was composed of 176 patients, after exclusion of patients with one or more of the following conditions: personal antecedents of cardiovascular or cerebrovascular events; body mass index under 18.5 kg/m² or greater than 35 kg/m²; significant alterations in muscle mass (amputations, loss of muscle mass, muscle diseases or paralysis); renal failure; pregnancy; lactation; severe hepatopathy; generalized edema; ascites; or incomplete anthropometric measurements or biochemical determinations.
We applied an instrument for collecting personal and biomedical data. The same interviewer always performed the interview to ensure standardization of the procedure. The participants underwent anthropometric-clinical measurements and a blood sample was taken. They were instructed to have a light dinner and to fast for 12 hours before blood collection. A partial morning urine sample was requested on the day when blood was collected.

Anthropometric, blood pressure and biochemical measurements
Weight and height measurements were made following standard protocols. Waist circumference was measured with a measuring tape at the midpoint between the last rib and the iliac crest, with the subject standing. This measurement was made at the end of an unstressed expiration. The waist circumference, body mass index (BMI; in kg/m 2 ) and waist-to-height ratio (WHR; in cm/cm) were classified as elevated in accordance with the accepted criteria. 10,11,12 Blood pressure was measured using a sphygmomanometer (Omron model M7; Omron Health Care, Kyoto, Japan). The diagnosis of arterial hypertension was established in accordance with international recommendations. 13 The percentage of body fat (%BF) was ascertained using a body composition analyzer (model TBF 300 A; Tanita, Tokyo, Japan). %BF ≥ 25% (men) and ≥ 30% (women) was considered elevated. 14 The A1 C hemoglobin fraction (HbA1 C ) in whole blood was assessed by means of an immunoassay. Glucose, creatinine, triglycerides (TGL), total cholesterol (TC), low-density lipoprotein-cholesterol (LDLc) and high-density lipoprotein-cholesterol (HDLc) were determined in serum using colorimetric enzymatic methods. Serum high-sensitivity C-reactive protein (hsCRP) was quantified by means of immunoturbidimetry. The protein content in the partial urine sample was determined using a reactive tape.
Detection of a protein level of at least one cross (+) in the urine was defined as proteinuria.
The estimated glomerular filtration rate (eGFR) was obtained through the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, 15 using the renal function calculator of the Spanish Society of Nephrology. 16 Presence of metabolic syndrome and its individual components were established in accordance with a harmonized definition. 11 Presence of diabetes was defined using the criterion of the American Diabetes Association. 17 TC, LDLc, TC/HDLc ratio, LDLc/ HDLc ratio and non-HDL cholesterol were classified as elevated in accordance with previously described criteria. 18,19,20,21 Existence of a proinflammatory state was defined as a hsCRP level ≥ 1 mg/l. In addition, the hsCRP level was classified as indicative of average cardiovascular risk when it was 1-3 mg/l or as indicative of high cardiovascular risk when it was ≥ 3.0 mg/l. 22 The level of glycemic control was categorized as "reduced" when HbA1 C was ≥ 5.7%; additionally, HbA1 C was categorized as < 5.7% (normal), 5.7%-6.4% (prediabetes) or ≥ 6.5% (diabetes). 17 eGFR was defined decreased using the cutoff points recommended through the guidelines of the National Kidney Foundation. 23

Continuous scores for cardiometabolic risk
The following continuous scores for cardiometabolic risk were evaluated: • Continuous metabolic syndrome severity Z-score (MetS Z-score): this was calculated by applying the equations proposed by Gurka et al. 24 for Hispanic individuals according to sex, using the calculator available at http://mets.health-outcomes-policy.ufl.edu/calculator/.
• Cardiometabolic index (ICMet): Wakabayashi and Daimon 25 proposed this index. It was calculated as the product of the TGL/HDLc ratio and WHR.
• Simple method for quantifying metabolic syndrome (siMS) score and siMS risk score): Soldatovic et al. 26 proposed these continuous scores. The first assesses the state of metabolic syndrome and the second evaluates the risk of coronary heart disease or cerebrovascular events. These scores were determined using the spreadsheet provided by Soldatovic et al. 26 and introducing the cutoff points of the metabolic syndrome definition applied in the present study.

Statistical analysis
Statistical Package for the Social Sciences (SPSS) software, ver- The variables studied were assessed with regard to normality of distribution, by means of the Kolmogorov-Smirnov test.
Variables that did not follow this distribution were transformed using the process described by Templeton Diabetes, metabolic syndrome and decreased glomerular function were more frequent among men (P < 0.05).
The medians for the siMS score, siMS risk score, ICMet and MetS Z-score were higher in patients with metabolic syndrome (P < 0.001) ( Figure 1A). The scores studied increased as the number of individual metabolic syndrome components also increased (P < 0.001) ( Figure 1B).
The variation of the scores according to the categories of the biomarkers assessed is shown in Table 2. All the scores were significantly higher in patients with elevated BMI, waist circumference, WHR, %BF, glucose, HbA1 C, TGL, TGL/HDLc ratio and hsCRP; all the scores were higher among patients with low HDLc. The siMS score, siMS risk score and MetS Z-score increased significantly as eGFR decreased; these indicators were also higher in patients with proteinuria. All the scores were significantly higher among diabetic patients. Only the siMS risk score was significantly higher among smokers and patients with a family history of cardiovascular disease; none of the scores was higher in hypertensive patients.
The linear regression analysis revealed that all the scores were positively correlated with BMI, waist circumference, WHR, %BF, glucose, HbA1 C , TGL, TC/HDLc, LDLc/HDLc ratio, TGL/HDLc ratio, non-HDL cholesterol, hsCRP and degree of proteinuria; all the scores were negatively correlated with HDLc. The siMS score, siMS risk score and MetS Z-score were inversely correlated with the eGFR ( Table 3). None of the scores studied showed correlations with LDLc after adjustment for age and sex.
The predictive value of the scores studied for the entire sample and according to sex are shown in Table 4 and Table 5. In the entire sample, MetS Z-score, siMS score and siMS risk score showed the ability to detect or discriminate metabolic syndrome, reduced glycemic control (HbA1 C ≥ 5.7%), proinflammatory state (hsCRP ≥ 1 mg/l) and decreased eGFR (< 90 ml/min/1.73 m 2 ); ICMet only had significant capacity to discriminate patients with metabolic syndrome and a proinflammatory state. Overall, the AUCs for MetS Z-score were significantly higher than the AUCs for the rest of the scores for discriminating metabolic syndrome, decreased glycemic control and proinflammatory state. Only the AUCs for MetS Z-score and siMS score for metabolic syndrome were similar. For reduced eGFR, the AUC for the siMS risk score was greater but did not differ significantly from the AUCs corresponding to siMS score and MetS Z-score.
Among women, all the scores assessed significantly discriminated metabolic syndrome and proinflammatory state. Only MetS Z-score had the capacity to detect reduced glycemic control, while siMS risk score showed the ability to discriminate reduced eGFR.
Among men, all the scores had predictive value for detecting the conditions studied.

DISCUSSION
The main purpose of this study was to examine the validity of four continuous scores that had been proposed for quantification of cardiometabolic risk. Overall, the four scores showed significant associations with most of the anthropometric and biochemical biomarkers that were measured. The scores studied showed predictive value for metabolic syndrome, reduced glycemic con- Z-score and risk factors, along with its ability to predict the progression of coronary heart disease and diabetes. 28,29,30,31 The MetS Z-score can also be highlighted as having the highest AUC for detecting three of the four conditions studied. In particular, it was the only score able to discriminate HbA1 C values ≥ 5.7% among women. The latter probably reflects the load factor that was obtained from glucose in constructing the equations for MetS Z-score, which was > 0.4 in women. 24 DeBoer et al. 32 corelated elevation of the MetS Z-score with declining eGFR, higher prevalence of microalbuminuria and higher incidence of chronic kidney disease in African-American women.
In our entire sample, MetS Z-score correlated negatively with eGFR Date expressed as mean ± standard deviation, median (interquartile range), n (%). Mann-Whitney U test. * P < 0.05 and ** P < 0.01 with respect to the first category of the biomarker. ‡ P < 0.05 and ‡ ‡ P < 0.01 with respect to the second category of the biomarker. BMI = body mass index; WC = waist circumference; WHR = waist to height ratio; SBP = systolic blood pressure; DBP = diastolic blood pressure; HbA1 C = A1 C hemoglobin fraction; TC = total cholesterol; LDLc = low-density lipoprotein cholesterol; HDLc = high-density lipoprotein cholesterol; TGL = triglycerides; hsCRP = ultrasensitive C-reactive protein; eGFR = estimated glomerular filtration rate. and positively with the degree of proteinuria. Nevertheless, it was only able to discriminate eGFR < 90 ml/min/1.73 m 2 in males.
This divergence may have been due to racial differences that affect susceptibility to deterioration of glomerular function and the distribution of the components of metabolic syndrome.
We assessed ICMet because it includes a few simple determinations that provide information on the metabolism of triglyceride-rich lipoproteins, insulin resistance and glycemic control.

Wakabayashi and Daimon 25 found a positive association between
ICMet and HbA1 C and showed that ICMet had significant predictive value for detecting diabetes and hyperglycemia (HbA1 C ≥ 5.7%) in Japanese women and men. Associations between ICMet and smoking habits, 33 progression of atheromatous plaque in patients with peripheral arterial disease 34 and the risk of hypertension 35 have also been reported.
In the present study, higher ICMet was observed in individuals with HbA1 C ≥ 5.7%, metabolic syndrome or diabetics. Likewise, ICMet correlated with most of the biomarkers that were measured, after adjustment for sex and age (except for systolic pressure, LDLc and degree of proteinuria). However, this measurement only had the capacity to detect metabolic syndrome and proinflammatory status in the entire sample and only showed predictive value for decreased glycemic control and reduced eGFR among men. ICMet also did not vary significantly among smokers or hypertensive patients. These observations place some doubt on the applicability of ICMet as a continuous measurement Table 3. Multiple linear regression analysis on continuous scores for cardiometabolic risk and biomarkers for cardiometabolic risk, inflammation and glomerular function, adjusted for age and sex

CONCLUSION
In a sample of Venezuelan adults, all the scores studied varied according to different anthropometric and biochemical biomarkers for cardiometabolic risk. They showed predictive value for metabolic syndrome and proinflammatory status. Three scores showed a predictive capacity regarding reduced glycemic control and decreased renal glomerular function. Because this study found certain differences in the performance of the scores studied, especially with regard to sex, selection of one or another will depend on the aim and the scope pursued. The aim in follow-up studies will be to confirm the present findings and their usefulness for prevention and intervention protocols relating to cardiometabolic diseases. Table 5. Area under the curve for continuous scores for cardiometabolic risk of detection of metabolic syndrome, reduced metabolic control, proinflammatory state and decreased estimated glomerular filtration rate, according to sex Data expressed as AUC (95% confidence interval). * P < 0.05 for the null hypothesis AUC = 0.05. ** P < 0.01 for the null hypothesis AUC = 0.05. *** P < 0.0001 for the null hypothesis AUC = 0.05. AUC = area under the curve; HbA1 C = A1 C hemoglobin fraction; hsCRP = ultrasensitive C-reactive protein; eGFR = estimated glomerular filtration rate.