de Síndrome Metabólica em Idosos : Uma Revisão Predictors of Metabolic Syndrome in the Elderly : A Review

O artigo teve como objetivo analisar criticamente estudos que avaliaram a habilidade dos indicadores antropométricos e clínicos enquanto capazes de predizer a SM em idosos. Foi realizado uma pesquisa bibliográfica nas bases de dados Medline/PubMed, LILACS e SciELO, além das referências de artigos selecionados e contato com autores. Foram analisados 21 artigos envolvendo indicadores antropométricos e clínicos em idosos, através de diferentes critérios da SM. Catorze estudos reportam aos indicadores antropométricos, sendo a circunferência da cintura (CC) e a razão cintura-estatura (RCEst) descritos como os melhores preditores de SM, com área sob a curva ROC (AUC) superior a 0,70 (p < 0,05). A circunferência do pescoço também foi descrita como indicador alternativo, porém com menor poder discriminatório. Para os indicadores clínicos, o lipid product accumulation (LAP) foi o parâmetro com melhor desempenho em identificar a SM em idosos, com AUC superior 0,85 e eficiência superior a 70%. Os indicadores CC, RCEst e LAP foram os mais sensiveis na predicação da SM. Desta forma, o emprego desses parâmetros podem facilitar a identificação precoce da SM através de métodos diagnósticos de fácil aplicação, boa precisão e baixo custo. Além disso, é importante a determinação de pontos de corte específicos para idosos, uma ver que a obesidade por si só, parece não ser forte preditor de SM em idosos.


Introduction
Metabolic syndrome (MetS) is defined as a set of risk factors that includes resistance to insulin, dyslipidemia, abdominal obesity and high blood pressure, and increases the risk of cardiovascular diseases and diabetes. 1,2The most up-to-date criteria to define MetS were prepared by the International Diabetes Federation (IDF) task force. 2 In their guidelines, it has been established that abdominal obesity is no longer a compulsory component, and specific cut-off points should be used to classify waist circumference (WC) by ethnic groups, in addition to criteria for changes in glucose and lipid metabolism and high blood pressure. 24][5] Clinical indicators are those which associate biochemical parameters to analysis measurements, particularly the lipid accumulation product (LAP) and visceral adiposity index (VAI). 3,4Anthropometric indicators include body mass index (BMI), WC, waist-to-hip ratio (WHR), waist-toheight ratio (WHtR), the sagittal abdominal indicator (SAD) and neck circumference (NC). 5,6e use of indicators to predict MetS may facilitate its identification in clinical practice, as they are simple, quick and functional.However, there is no consensus on the best indicator able to identify MetS in the elderly due to different functional characteristics and varied cut-off points, many of which are specific to young adults, and with different criteria for defining MetS. 1,2Nevertheless, early identification of MetS is important in this age group since it may assist health teams to decide on strategies aimed at reducing global cardiovascular risk. 1 using the following word combinations: ("metabolic syndrome" or "syndrome X" or "plurimetabolic syndrome") and ("elderly" or "older adults" or "aged") and ("predict" or "identify" or "ability").These expressions were searched either in combination with each other or alone.Articles of interest listed in the references were also identified and reviewed.Several authors were contacted for relevant information not provided in the articles.

Criteria for including the articles
Only original articles written in English, Spanish or Portuguese were included in the review.Articles published between January 2010 and January 2016 were selected and classic studies published on the topic prior to this period were also included.Population-based studies in the elderly or studies on institutionalised elderly people that evaluated anthropometric and clinical indicators (LAP, VAI or lipid ratios) as predictors of MetS were included.

Data extraction
After screening of titles and abstracts according to eligibility criteria, relevant articles were selected for fulltext reading (Figure 1).
The following data were extracted independently: 1) characteristics of the article (authors, journal and year of publication); 2) location where the study was performed (city and country); 3) characteristics of the population studied (number of participants, sex and age range);

Results
Two hundred and eighty-nine studies were identified on the bibliographic databases searched.After reading the titles and abstracts, 244 studies were excluded.Thus, 45 articles were eligible for evaluation, and 26 of these papers were excluded as: 7 presented data on a young adult population; 5 evaluated global cardiovascular risk; 4 evaluated isolated MetS components (dyslipidemia, high blood pressure, diabetes or obesity); 2 evaluated predictors of resistance to insulin; 5 evaluated complete blood count components as predictors of MetS; 1 were with patients admitted to hospital and 2 studies written in Chinese.Finally, after reviewing the bibliographic references for articles of interest, two further papers were included.Thus, 21 articles were analysed in this review (Figure 1).
Of the articles analysed, 13 discussed anthropometric indicators (Tables 1 and 2), 7 evaluated clinical indicators (LAP, VAI and lipid ratios) (Tables 3 and 4) and one  To be continued article 7 compared anthropometric indicators and lipid ratios to predict MetS in the elderly (and hence is cited in anthropometric and clinical indicators' tables).
The IDF's and the NCEP-ATP III's criteria were most commonly used criteria for definition of MetS).][9][10] The harmonized criteria states that obesity itself is not a prerequisite for diagnosis of MetS, which should be established based on the presence of any 3 or 5 risk factors (among which obesity is included).

Anthropometric indicators
Fourteen studies reported anthropometric indicators as predictors of MetS (Tables 1 and 2).Among the anthropometric indicators, those most cited were BMI, WC and WHR.The most recent studies analysed WHtR, SAD and NC.
The majority of the papers 5,7,[11][12][13] which evaluated anthropometric indicators highlight WC and WHtR as the best predictors of MetS in the elderly, when compared to BMI and WHR.The works of Liang et al. 7 and Guasch-Ferré et al. 9 reported an AUC over 0.70 for WC and WHtR to detect MetS, considering the harmonized criteria for MetS.For women, Liang et al. 7 identified the cut-off points of 79.5 cm for WC (sens: 72.7%; spec: 76.7%) and 0.53 for WHtR (sens: 67.6%; spec: 72.9%), with efficiency higher than 70%.Zeng et al,. 13evaluating the presence of a minimum of two MetS components and stratifying the results by sex and age, observed similar AUCs and cut-off points for elderly women at the age range of 60-69 and 70-79 years (Table 2).
Among the indicators with unsatisfactory performance, Paula et al. 14 established that the BMI and the sum of four skinfolds were the anthropometric parameters of adiposity which presented the least efficiency in identifying MetS.
Another indicator cited was SAD, which is able to estimate the excess of visceral fat and, consequently, is a better predictor of cardiometabolic risk than classic indicators. 15Only two studies 16,17 evaluated SAD in the elderly; Sharda et al. 17 aimed to identify the best SAD cut-off points to predict MetS, and suggested the cutoff points of 22 cm for men and 20 cm for women, with sensitivity and specificity higher than 80% for both sexes.Similar cut-offs were found by Risérus et al. 16 using a cardiometabolic risk in elderly subjects.
In relation to NC, the articles analysed its performance in identifying MetS and its isolated components.Aoi et al. 18 showed that the increase in NC is associated to a rise in metabolic risk factors, such as resistance to insulin (leptin, HbA1c and HOMA-IR) and lipid profile (TG and HDL).
In the studies 8,19,20 which evaluated the predictive capacity of the NC, the cut-off points suggested by the authors were quite similar, varying between 35 and 41 cm for men and 33 and 35 cm for women (Table 2).Hoebel et al. 20 observed that white men presented a higher cut-off point (41 cm) when compared to black men (35 cm).The cut-off points for women were similar between white and black women, although the cut-off point for black women did not predict MetS, in contrast to what was observed for men.Limpawattana et al.   and Yan et al. 8 presented similar cut-off points for NC to predict MetS in individuals aged over 50 and 65 respectively, and Limpawattana et al. 19 showed an efficiency over 72% for the cut-off point suggested for men (sens: 70.89%; spec: 81.15%) and women (sens: 86.54%; spec: 59.39%).
A parameter that has been insufficiently investigated was the Fat Mass Index (FMI), which is normalizes body fat for individual's height.This index has been suggested to independently evaluate body fat from changes in fat free mass.In the study performed by Liu et al., 21 FMI was the parameter with the highest AUC, and high FMI showed significantly higher odds ratio for MetS than the low FMI in both sexes.These authors suggested that a FMI of 7.00 kg/m² for men and 7.90 kg/m² for women, evaluated by bioelectrical impedance analysis, could predict the presence of MetS.

Clinical indicators
Tables 3 and 4 demonstrate the characteristics and results, respectively, of the eight studies which evaluated LAP and VAI clinical indicators and lipid ratios as predictors of MetS according to NCEP and IDF criteria and harmonized criteria for MetS.
LAP is an indicator proposed to estimate lipid concentration in adults, described for the first time by Kahn. 3 It is an emerging cardiovascular risk index based on the product of WC multiplied by fasting triglyceride concentration.LAP has proven to be a reliable marker for cardiovascular disease in adults, outperforms other predictors of this risk as BMI. 3 Most of the studies analysed evaluated a mixed population, stratified by sex and/or age range.The cut-off point for LAP suggested by the authors [22][23][24][25][26] varied from 31.6 to 51.8, with a difference between the sexes (Table 4).Among men, the LAP AUC to predict MetS varied from 0.84 to 0.937, while for women the lowest AUC was 0.78 and highest was 0.904 (Table 4).
Taverna et al. 23 reported a lower AUC for men and women in the 61-70 age range, considering the IDF criteria (area under the curve, AUC: 0.84; 95% CI: 0.74-0.93 for men and 0.78; 95% CI: 0.76-0.99 for women) when compared to the other studies.
Ejike CECC 25 investigated LAP as a predictor of MetS in the elderly (aged between 65 and 84); the authors reported an AUC of 0.937 (p=0.013) and suggested a LAP cut-off point of 43.9 (sens: 100%; spec: 81%).Tellechea et al. 22 , Chiang & Koo 24 and Motamed et al. 26 also observed the high predictive capacity of LAP to identify MetS (AUC higher than 0.90).Among these studies, the lowest LAP cut-off point to identify MetS was reported by Chiang & Koo 24 (31.6 for both sexes).
VAI is a gender-specific, empirical-mathematical model based on a combination of anthropometric measurements (WC and BMI) with biochemical parameters (TG and HDLc), which may be and indicator of fat distribution and function. 4The Alkam Metabolic Syndrome Study 4 introduced VAI as a new marker of adipose tissue dysfunction, independently associated with cardiovascular events, which was not observed for WC and BMI.Only two studies investigated this indicator in MetS. 25,27For Amato et al. 27 , a VAI value higher than 2.00 (sens: 68.5% and spec: 76.0%) for individuals aged over 66 was able to predict MetS.However, in the study by Ejike CECC, 25 VAI showed an AUC of 0.640 but without statistical significance (p = 0.426), suggesting a cut-off point higher than 4.4, with relatively better sensitivity and specificity to identify MetS in the elderly (sens: 67% and spec: 84%).Two papers 7,10 described the results of lipid ratios, with different cut-off points between the studies.Arthur et al. 10 observed that the TG/HDL and HDL/CT ratios presented AUC of 0.80 (95% IC: 0.70-0.90)for both ratios, suggesting a cut-off point of 0.61 (sens: 87.2% and spec: 80%) for the TG/HDL ratio and 0.34 (sens: 96.6% and spec: 83.3%) for the HDL/CT ratio, therefore demonstrating efficiency of over 83%.

Discussion
The anthropometric indicators that showed a better performance in identifying MetS were WC and WHtR (among the classic indicators) and NC (among the recently studied indicators).Regarding the clinical parameters, LAP was the best indicator, followed by the HDL/CT ratio.
An important point to be discussed is the criteria used to define MetS (IDF, NCEP ATPIII and harmonized criteria), which may justify the diversity of results discussed in this review.In addition, the use of different criteria and measurement techniques, particularly in relation to WC, does not ensure comparability between the studies.WC and WHtR were the best parameters evaluated, 5,7,9,13 with WC showing higher sensitivity among the obesity indexes.In addition, it was observed that the WC cut-off points suggested were lower than those proposed by the MetS criteria used, revealing that non-specific cut-off points for ethnic and age groups can underestimate the presence of MetS, depending on the criterion used for its classification.However, Chu et al. 11 observed that WC and WHtR showed similar results to BMI and WHR, with an AUC lower than 0.70 to detect a minimum of three MetS criteria in women.This divergence could be attributed to the age of the elderly (average of 71.9 years), with high prevalence of other comorbidities, apart from obesity, that contribute to the development of MetS.Therefore, obesity appears not to be a decisive criterion for MetS detection in this age range.
A further indicator discussed in the literature is SAD.Although very few studies have evaluated this indicator as a predictor of MetS, the cut-off points suggested by Risérus et al. 16 and Sharda et al. 17 are in line with other studies that evaluated SAD as a predictor of visceral fat. 15,28Despite being related to visceral abdominal fat deposits, SAD has not been widely used in clinical practice since it requires a specific measuring instrument (abdominal caliper -Holtain, Ltd, Dyfed, Wales, U.K).
NC was also evaluated, with similar cut-off points suggested by the authors, 8,19,20 which can be explained by the fact that this measurement is little affected by agingrelated changes in body composition.The association between NC and MetS may be related to the fact that subcutaneous fat in the neck area is responsible for higher systemic release of free fatty acids. 19This excess of free fatty acids is associated with resistance to insulin, hypertriglyceridemia, vascular lesion and high blood pressure, and could predict MetS.
Differences between sexes and ethnic groups were reported in the study by Hoebel et al. 20 The authors observed that a NC cut-off point of 35 cm was not able to predict MetS in black women, in contrast with what was observed in men, suggesting a concept of "healthy obesity".However, a limitation of this study was that the authors considered individuals aged between 46 and 65 as the elderly group and, for this reason, their results cannot be extrapolated to individuals aged over 65.Altogether, the studies 8,[18][19][20] suggested the use of NC, as it is a simple and practical anthropometric parameter, able to identify MetS.Another advantage of the NC is the fact that it is not affected during respiratory movement or in the postprandial period.
FMI was only discussed in the study by Liu et al. 21lthough the authors suggested a cut-off point, it is worth pointing out that there was no stratification by age group in the data analysis.In addition, the study was conducted in China, in a smaller population with brevilineal biotype, which limit the extrapolation of results to other populations.
The studies included showed that the cut-off points for obesity indicators need to be specific for the elderly, with higher values when compared to other age groups, as obesity does not appear to be a strong predictor of MetS in the elderly.
With regards to clinical indicators, LAP was the parameter most discussed in the studies.Kahn 3 highlights that the WC and TG levels tend to increase with age, accumulating over time.From this perspective and considering that lipid accumulation intensifies cardiometabolic consequences, the use of LAP seems an advantageous approach in evaluating MetS and cardiovascular risk.
The cut-off points for LAP suggested by the studied showed a wide variety, which could be justified by ethnic diversity, extensive age ranges and different criteria used to define MetS.In addition, two studies 22,26 did not adequately define the age group of the elderly.While Motamed et al. 26 did not stratify the results between adults and the elderly, Tellechea et al. 22 considered the elderly group to be aged between 45 and 65.These facts limit the understanding of this indicator's performance, specifically among the elderly.
Although Taverna et al. 23 has reported the lowest LAP AUC for the elderly aged between 61-70 years when compared with other studies, these authors show a better performance of LAP as a predictor of MetS in the elderly.Besides, although Ejike CECC 25 demonstrated good results using the LAP indicator, a small sample (only 40 men) was included, which was an important limitation of the study.
The lowest LAP cut-off point to predict MetS suggested by Chiang & Koo, 24 could be related to the criterion used to define MetS and the physical characteristics of the population evaluated.The capacity of LAP to identify MetS, compared with that of other parameters, is associated to the fact that this indicator reflects anatomical and physiological alterations associated to visceral fat and its accumulation, 22,25 indicating that, regardless of how excess lipids are stored, it will be identified by LAP. 25 Another clinical indicator also related to the dysfunction of the adipose tissue is VAI, which uses WC, BMI, TG and LDL measurements for its calculation.Only two studies 25,27 have analysed its performance and proposed a VAI cut-off point to identify MetS.However, the cut-off points suggested diverged among the authors.It is of note that three of the variables on which VAI is based (WC, TG and HDL) are also MetS components, which makes the use of this index in patients with clear MetS pointless. 29The studies 7,10 using lipid ratios showed that the use of this indicator in isolation is not able to identify MetS, particularly among the elderly, but rather to identify isolated cardiovascular risk factors.
The results were clear in showing that WC, WHtR and LAP indicators were the most sensitive ones for predicting MetS.Therefore, the use of these parameters may facilitate the early identification of MetS through easily-applied, precise and low cost diagnostic methods.The use of these indicators by health professionals is important to optimise MetS prevention and treatment, and to minimize its complications.
To summarise, the studies included propose different indicators to predict MetS.However, the cut-off points suggested vary between each other, as they are linked to different populations, with different age and ethnic characteristics.Further studies are required to identify the best practical, validated method to predict MetS.Also, studies comparing different indicators are needed, considering specific aspects, such as sex, well-defined age ranges and ethnic groups.
Different MetS classification criteria have been adopted, which makes the comparison between the studies problematic and the standardisation of the criteria essential.In addition, there is no consensus on the best indicator to predict MetS in the elderly, and studies in this population is still scare.These data affect the understanding of the indicators' performance in predicting MetS in this age group.
The most of these studies presented to methodological differences and study designs, as well as limitations inherent to observational studies, which make the comparison between results and the control of population biases difficult.

Conclusion
Analysis of the studies included in this review allowed us to infer that the WC, WHtR and LAP indicators presented the best predictive power for MetS in the elderly because they are strongly related to visceral abdominal fat deposits.However, the disparate values between the studies substantiated that the prevalence of other comorbidities, rather than obesity, may be contributing to the development of MetS in this population.Thus, the results revealed that the distribution of abdominal fat, instead of overweight appears to be a fundamental criterion for this age group, but the distribuition of abdominal fat.The need to identify a simple method which facilitates the approach and detection of high risk for MetS in the elderly is highlighted, allowing the early prevention of cardiovascular events.

Figure 1 -
Figure 1 -Flowchart of selection of papers for the review.

Table 1 -Studies evaluating the anthropometric indicators as predictors of metabolic syndrome in the elderly Authors and publication year (Ref. nº.)
* plane the midpoint between the lower costal rib and the iliac crest, perpendicular to the long axis of the trunk.†at the level of the umbilicus.

Table 3 -Studies evaluating the clinical indicators as predictors of metabolic syndrome in the elderly Authors and publication year (Ref. nº.)
W: women; M: men; MetS: metabolic syndrome IDF: international diabetes federation; NCEP ATPIII: National Cholesterol Education Programme Adult Treatment Panel III; LAP: lipid accumulation product; VAI: visceral adiposity index; WC: waist circumference; WHR: waist-to-hip ratio; WHtR: waist-to-height ratio.BMI: body mass index; HDLc/TC: HDL-cholesterol to total cholesterol ratio; HDL/LDL:HDL-cholesterol to LDL: cholesterol ratio; TG/LDL: triglyceride to HDL: cholesterol ratio.