Concordance between Different Criteria for Metabolic Syndrome in Peruvian Adults Undergoing Bariatric Surgery

Background: Metabolic Syndrome (MetS) is a clinical entity that has been linked to several non-communicable diseases. There are various consensuses to determine its presence, such as the IDF, ALAD, Harmonized, AHA/NHLBI, NCEP-ATP III or AACE criteria. However, there is currently no standardization to properly identify it. Objective: To assess the diagnostic concordance between different criteria for MetS in Peruvian adults undergoing bariatric surgery. Methods: We conducted a secondary analysis of the institutional database of a bariatric clinic located in Lima, Peru. We obtained data from adults between 18–59 years who underwent bariatric surgery (Roux-en-Y Gastric Bypass or Sleeve Gastrectomy). According to the Kappa coefficient, a heatplot was designed to analyze the concordance of the criteria. Results: An almost perfect concordance was found between all criteria except AACE. The highest kappa coefficient (κ = 0.980) was recorded between the IDF and ALAD criteria using all the sample. Similar results were obtained when we stratified by sex. Conclusions: This study shows that, excluding the AACE, different criteria for metabolic syndrome could be used in Latino adults undergoing bariatric surgery with similar results. Given the postoperative implications, we believe that IDF and ALAD would be the best options in our population.


Introduction
Metabolic Syndrome (MetS) is an entity that comprises multiple abnormal metabolic conditions [1] linked with increased cardiovascular risk [2]. Some related diseases include coronary heart disease, type 2 diabetes mellitus (T2DM), polycystic ovary syndrome, nonalcoholic fatty liver (NAFLD), asthma, sleep disorders and some types of cancer [3,4]. Approximately a quarter of world s population has MetS [5].
It is essential to properly identify patients with MetS in order to provide early treatment and prevent unwanted clinical outcomes [6]. The treatment of MetS includes lifestyle changes, pharmacological management and surgical therapies such as bariatric surgery (BS) [1]. BS causes weight loss by different procedures that include gastric volume restriction and malabsorption [7]. Hence, as one of the objectives of BS is the resolution of MetS, it is essential to reach a standardized definition in this population [8].
Multiple studies have evaluated MetS in people who have undergone BS, each with different diagnostic criteria [9][10][11]. In recent years, there has been increased use of the International Diabetes Federation (IDF) definition for MetS, especially in the European adult population [8,12]. Other studies in Hispanic adults have used the definition provided by the Harmonized criteria, in which the elevated waist circumference (WC) is replaced by a body mass index (BMI) ≥ 30 kg/m 2 [13]. Finally, few studies use the definition provided by the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP-III), 2 of 9 research conducted in the United States population, which uses a highest cut-off value for WC as a necessary factor [14,15].
The Latino population has epidemiological and genetic factors linked to elevated MetS incidence. These include ethnicity, malnutrition during childhood, maternal malnutrition, high carbohydrate diet, aging population and a high prevalence of insulin resistance [16,17]. Latin America reports a prevalence of MetS of 25.3% for women and 23.2% for men, higher than other regions [17][18][19]. In Peru, cross-sectional studies using International Diabetes Federation (IDF) criteria report a prevalence of approximately 30% [20,21]. Due to the lack of epidemiological data and differences in criteria used, the consensus of the Latin American Diabetes Association (ALAD, by its acronym in Spanish) was carried out in 2010 and proposed five clinical and laboratory values for the diagnosis of MetS [16], based on diverse epidemiologic and clinical factors recommended by reviewing published literature from Latin American countries [16]. However, to date, there are no studies evaluating concordance between ALAD and other criteria for MetS [22,23].
This study aimed to evaluate the diagnostic concordance between six different criteria for MetS in Peruvian adults undergoing bariatric surgery.

Study Design and Population
This was a secondary analysis that used the institutional database from a bariatric clinic located in Lima, Peru. We included data from 18-59 years old adults who underwent bariatric surgery (either Roux-en-Y Gastric Bypass (RYGB) or Sleeve Gastrectomy (SG)). Hence, we included: (a) patients with BMI ≥ 30 kg/m 2 who had unsuccessfully tried to lose weight with lifestyles or pharmacological methods and T2DM and/or poorly controlled arterial hypertension in spite of optimal medical therapy; (b) patients with BMI ≥ 35 kg/m 2 who had unsuccessfully tried to lose weight with lifestyles or pharmacological methods and T2DM, hypertension, sleep apnea NAFLD and/or other poorly controlled comorbidities secondary to obesity in spite of optimal medical therapy; (c) patients with BMI ≥ 40 kg/m 2 , and; (d) patients (selected on a case-by-case basis) with NAFLD and/or insulin resistance with an inability to achieve a healthy weight loss sustained for a period of time with prior non-surgical weight loss methods.

Criteria for Metabolic Syndrome
We considered six different criteria for MetS that were the most commonly used in the current literature (Table 1): Latin American Consensus of the Latin American Diabetes Association (ALAD): MetS was defined by the presence of abdominal obesity: WC ≥ 94 cm (men) and ≥88 cm (women), plus two of the following four conditions: Triglycerides (TG) > 150 mg/dL or on hypertriglyceridemia treatment, High-density lipoprotein cholesterol (HDL-C) < 40 mg/dL (men) and < 50 mg/dL (women) or on treatment for low HDL-C, high blood pressure (BP) as systolic blood pressure (SBP) ≥ 130 mmHg and/or diastolic blood pressure (DBP) ≥ 85 mmHg or on antihypertensive drugs, impaired fasting glucose (IFG)/impaired glucose tolerance (IGT) or diabetes [16].
Harmonized criteria: MetS was defined by the presence of three of these five conditions: increased WC depending on the population/country, TG ≥ 150 mg/dL or on hypertriglyceridemia treatment, HDL-C < 40 mg/dL (men) and <50 mg/dL (women) or on treatment for low HDL-C, SBP ≥ 130 mmHg and/or DBP ≥ 85 mmHg or on antihypertensive treatment, fasting plasmatic glucose (FGP) ≥ 100 mg/dL or on hyperglycemia treatment [24].
International Diabetes Federation (IDF): MetS was defined by a WC ≥ 90 cm for males and WC ≥ 80 for females plus at least two of the following four conditions: TG > 150 mg/dL or on treatment for this lipid abnormality, HDL-C < 40 mg/dL (men) and <50 mg/dL (women) or on treatment for reduced HDL-C, SBP ≥ 130 and/or DBP ≥ 85 mmHg or on treatment for hypertension, raised FPG ≥ 100 mg/dL or previously diagnosed T2DM [25].
American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement (AHA/NHLB): MetS was defined by the presence of three of these five conditions: WC ≥ 102 cm (men) and ≥ 88 cm (women), TG ≥ 150 mg/dL or on hypertriglyceridemia treatment, HDL-C < 40 mg/dL (men) and < 50 mg/dL (women) or on treatment for reduced HDL-C, SBP ≥ 130 mm Hg and/or DBP ≥ 85 mm Hg or on antihypertensive treatment, fasting glucose ≥ 100 mg/dL or on hyperglycemia treatment [4].
American Association of Clinical Endocrinologists (AACE): Clinical definition of metabolic syndrome depends on clinical judgment according to the presence of BMI ≥ 25 kg/m 2 , TG ≥ 150 mg/dL and HDL-C < 40 mg/dL (men) and < 50 mg/dL (women), BP ≥ 130/85 mmHg or antihypertensive treatment, 2-h post glucose challenge > 140 mg/dL, Fasting glucose 110-126 mg/dL. The AACE criteria does not considerate treatment for specific diseases except for hypertension. Additional risk factors include a family history of T2DM, hypertension, coronary heart disease (CHD), polycystic ovary syndrome (PCOS), sedentary lifestyle, advanced age, ethnic groups at high risk for T2DM or for CHD [26].

Statistical Analysis
We presented categorical data as frequencies and percentages (%) and numerical data with mean and standard deviation (SD) or median and interquartile range (IQR), according to the distribution. Student s t-test or Mann-Whitney U test was used to assess significant differences between baseline characteristics according to sex.
Kappa (k) statistic was used for assessing the agreement between the different criteria of MetS and a heatplot was designed. p-values < 0.05 were considered significant, and all the analyses were performed using the statistical package Stata v15.1 (StataCorp, College Station, TX, USA).

Ethics
The present study was approved by the Institutional Review Board of the Clínica Avendaño. Participant consent was not required, and the study database was de-identified.

Characteristics of Participants
In total, 205 participants met the inclusion criteria and were enrolled in this study. The 62.9% (n = 127) were women and the median BMI was 37.44 kg/m 2 (140 and 65 patients had a BMI < 40 kg/m 2 and ≥40 kg/m 2 , respectively). In addition, 15 (7.3%), 37 (18.1%) and 10 (4.9%) reported a history of T2DM, hypertension, and obstructive sleep apnea that had not responded effectively to medical therapy. NAFLD and insulin resistance were present in almost all (>90%) of the participants.

Concordance of Metabolic Syndrome Definitions
In Figure 1, we presented a heatplot designed to assess the concordance between the different MetS criteria. An almost perfect concordance was found between all criteria except the AACE criteria. The highest kappa coefficients were recorded between the IDF and ALAD (κ = 0.980), AHA/NHLBI and Harmonized criteria (κ = 0.971), and AHA/NHLBI and NCEP-ATP III (κ = 0.932). The criteria that showed the least concordance with the others was AACE.

Concordance of Metabolic Syndrome Definitions
In Figure 1, we presented a heatplot designed to assess the concordance between the different MetS criteria. An almost perfect concordance was found between all criteria except the AACE criteria. The highest kappa coefficients were recorded between the IDF and ALAD (κ = 0.980), AHA/NHLBI and Harmonized criteria (κ = 0.971), and AHA/NHLBI and NCEP-ATP III (κ = 0.932). The criteria that showed the least concordance with the others was AACE. In Figure 2, we present a heatplot of MetS diagnostic criteria in men. A perfect concordance was found between IDF and ALAD (κ = 1). An almost perfect concordance was found between AHA and Harmonized (κ = 0.971); AHA and NCEP-ATP III (κ = 0.916), and; Harmonized with ALAD and IDF (κ = 0.941). Similar to Figure 1, AACE criteria showed the weakest concordance with other criteria.
In Figure 2, we present a heatplot of MetS diagnostic criteria in men. A perfect concordance was found between IDF and ALAD (κ = 1). An almost perfect concordance was found between AHA and Harmonized (κ = 0.971); AHA and NCEP-ATP III (κ = 0.916), and; Harmonized with ALAD and IDF (κ = 0.941). Similar to Figure 1, AACE criteria showed the weakest concordance with other criteria.

Main Findings
This study was conducted in adults undergoing bariatric surgery to assess the con-

Main Findings
This study was conducted in adults undergoing bariatric surgery to assess the concordance between the criteria for metabolic syndrome. We found an almost perfect concordance between IDF/ALAD/Harmonized/AHA/NHLBI/NCEP-ATP III criteria, regardless of the sex of the study subject. The highest concordance in the study was found between ALAD and IDF.

Comparison with Other Studies
The lowest prevalence of metabolic syndrome was found with the AACE criteria (22.9%) and the highest prevalence was found with the IDF criteria (59.5%). In Peru, a previous study among adults with overweight and obesity reported the prevalence of MetS using different criteria. The lowest prevalence of MetS in this study was found with the WHO criteria (42%) and the highest prevalence was found with the Szabo criteria (74.3%). Comparing their results with ours, according to the criteria, we found the following prevalence of MetS: IDF (58.6% vs. 59.5%), AHA / NHLBI (52.9 % vs. 54.1%) and NCP-ATP (56% vs. 50.7%) [23]. Another study conducted by Vasquez MA et al. (2016) in Ecuador, found that the concordance between IDF and ALAD criteria was almost perfect, as in our study [27].
The most frequent criteria used in epidemiological and clinical studies in Peru are IDF, ALAD and NCP-ATP III [20,21,28]. In our study, we found an almost perfect concordance between these criteria. A similar concordance was founded in a previous study conducted in Peruvian population, although they did not consider the ALAD criteria [23].

Results Interpretation
Current research confirmed the concordance between ALAD and IDF criteria, possibly due to the indispensable use of the waist circumference as criterion, unlike the other criteria. Likewise, the cut-off points of the waist circumference are relatively similar, mainly in men [16,25]. In our study, AACE was the one with the weakest level of concordance with the others, possibly due to the fact that IFG or IGT are indispensable or that AACE does not considered actual treatment for lipid abnormalities or hyperglycemia [26]. Our database had fasting glucose values; however, we do not have values from the oral glucose-tolerance test; so we cannot detect if the patient has IGT. Therefore, it could not be correctly assessed whether patients meet the AACE criteria.

Relevance in Public Health and Clinical Practice
Metabolic syndrome is characterized by a progressive deterioration of quality of life and a strong association with multiple diseases such as T2DM, hypertension, dyslipidemia and obesity [29]. Moreover, less labor productivity and absenteeism associated with this condition contribute to a detriment of the family and personal economy [29]. The prevalence of MetS is increasing worldwide, especially in recent years [30]. In Latin America, the prevalence of MetS varies by city, the lowest prevalence is reported in Quito (13.7%) and the highest in Mexico City (27.0%) [31]. We did not find national representative data from Peru.
As previously mentioned, there is high concordance between the evaluated criteria, with the exception of the ACEE criteria. Thus, in the adult population undergoing bariatric surgery and, by extrapolation, in the population of adults with obesity/overweight, could make sense to replace one criteria of metabolic syndrome with another in the case of not being able to use one of them. This recommendation is mainly between the IDF and ALAD, AHA and Harmonized, and AHA and NCEP-ATP III criteria. Specifically, since adults with MetS have an increased risk of morbidity and mortality after bariatric surgery [14], sufficiently sensitive criteria should be used for their accurate diagnosis. In this sense, ALAD and IDF could be the best choices for Latin American adults undergoing bariatric procedures.

Limitations
The study had some limitations. First, the study cannot be extrapolated to adolescents or older adults since only participants between 18-59 years of age who underwent bariatric surgery were considered; however, we consider that our results can be extrapolated to the population of adults with overweight/obesity who are candidates for bariatric surgery. Second, there is a possibility that some variables were mismeasured. However, the personnel that filled the database received intensive training on the correct filling of the base. We also conducted a rigorous assessment of the data quality, which consisted of identifying missing and implausible data, in addition to independent double coding and cross-checking of the databases, in order to reduce the possibility of information bias. Third, the absence of variables such as oral glucose tolerance, family history of DM2 and polycystic ovary syndrome might have limited the assessment of the AACE criteria.

Conclusions
This study shows that excluding the AACE, different criteria for metabolic syndrome could be used in Latino adults undergoing bariatric surgery with similar results. Given the postoperative implications, we believe that IDF and ALAD would be the best options for our population, although further studies might be necessary to extrapolate our results to similar populations from other Latin American countries. Similarly, future cohort design studies should compare the concordance of these criteria in the middle-and long-term follow-up (i.e., assessing the resolution of MetS after bariatric surgery).