<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.1 20151215//EN" "https://jats.nlm.nih.gov/publishing/1.1/JATS-journalpublishing1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink" specific-use="sps-1.9" dtd-version="1.1" xml:lang="pt" article-type="research-article">
  <front>
    <journal-meta>
      <journal-id journal-id-type="publisher-id">abc</journal-id>
      <journal-title-group>
        <journal-title>Arquivos Brasileiros de Cardiologia</journal-title>
        <abbrev-journal-title abbrev-type="publisher">Arq. Bras. Cardiol.</abbrev-journal-title>
      </journal-title-group>
      <issn pub-type="ppub">0066-782X</issn>
      <issn pub-type="epub">1678-4170</issn>
      <publisher>
        <publisher-name>Sociedade Brasileira de Cardiologia - SBC</publisher-name>
        <publisher-loc>S&#227;o Paulo, SP, Brazil</publisher-loc>
      </publisher>
    </journal-meta>
    <article-meta>
      <article-id pub-id-type="publisher-id" specific-use="scielo-v3">sQqCcvyf4wRRdftGVx4mh9G</article-id>
      <article-id pub-id-type="other">00006</article-id>
      <article-id pub-id-type="publisher-id" specific-use="scielo-v2">S0066-782X2005001400006</article-id>
      <article-id pub-id-type="publisher-id" specific-use="scielo-v1">S0066-782X(05)08500106</article-id>
      <article-id pub-id-type="doi">10.1590/S0066-782X2005001400006</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Artigos Originais</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Indicadores antropom&#233;tricos de obesidade como instrumento de triagem para risco coronariano elevado em adultos na cidade de Salvador - Bahia</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <name>
            <surname>Pitanga</surname>
            <given-names>Francisco Jos&#233; Gondim</given-names>
          </name>
          <xref ref-type="aff" rid="aff01"/>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Lessa</surname>
            <given-names>Ines</given-names>
          </name>
          <xref ref-type="aff" rid="aff01"/>
        </contrib>
      </contrib-group>
      <aff id="aff01">
        <addr-line>
          <named-content content-type="city">Salvador</named-content>
          <named-content content-type="state"/>
        </addr-line>
        <institution content-type="orgname">Universidade Federal da Bahia</institution>
        <country country="BR">Brazil</country>
      </aff>
      <pub-date date-type="collection" publication-format="electronic">
        <month>07</month>
        <year>2005</year>
      </pub-date>
      <volume>85</volume>
      <issue>1</issue>
      <fpage>26</fpage>
      <lpage>31</lpage>
      <history>
        <date date-type="received">
          <day>11</day>
          <month>05</month>
          <year>2004</year>
        </date>
        <date date-type="accepted">
          <day>21</day>
          <month>01</month>
          <year>2005</year>
        </date>
      </history>
      <permissions>
        <license license-type="open-access" xlink:href="http://creativecommons.org/licenses/by-nc/4.0/" xml:lang="en">
          <license-p>This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.</license-p>
        </license>
      </permissions>
      <abstract>
        <p>OBJETIVO: Comparar v&#225;rios indicadores antropom&#233;tricos de obesidade e identificar dentre eles qual melhor discrimina o risco coronariano elevado (RCE). M&#201;TODOS: Estudo de corte transversal, com amostra composta por 968 adultos de 30 a 74 anos de idade, sendo 391 (40,4%) do sexo masculino. Foram constru&#237;das diversas curvas Receiver Operating Characteristic (ROC) e comparadas &#224;s &#225;reas sob as mesmas entre o &#237;ndice de conicidade (&#237;ndice C), &#237;ndice de massa corporal (IMC), raz&#227;o circunfer&#234;ncia cintura-quadril (RCCQ), circunfer&#234;ncia de cintura (CC) e RCE. Verificou-se tamb&#233;m a sensibilidade e especificidade para identificar e comparar o melhor ponto de corte entre os diversos indicadores de obesidade para discriminar o RCE. Foi utilizado intervalo de confian&#231;a a 95%. RESULTADOS: A maior &#225;rea sob a curva ROC foi encontrada entre o &#237;ndice C e RCE, em indiv&#237;duos do sexo masculino, 0,80 (0,74-0,85), diferindo significativamente dos demais indicadores de obesidade. Em mulheres, a maior &#225;rea sob a curva ROC encontrada foi de 0,76 (0,71-0,81), sendo iguais entre &#237;ndice C, RCCQ e RCE. CONCLUS&#195;O: Esses resultados demonstram que o &#237;ndice C e RCCQ s&#227;o os melhores indicadores de obesidade para discriminar RCE. A CC tem intermedi&#225;rio poder discriminat&#243;rio e o IMC foi o indicador antropom&#233;trico de obesidade menos adequado para discriminar RCE. Estes dados sugerem que os indicadores de obesidade abdominal s&#227;o melhores para discriminar RCE que os indicadores de obesidade generalizada.</p>
      </abstract>
      <kwd-group xml:lang="pt" kwd-group-type="author-generated">
        <kwd>indicadores antropom&#233;tricos de obesidade</kwd>
        <kwd>risco coronariano</kwd>
        <kwd>obesidade abdominal</kwd>
      </kwd-group>
      <pub-date publication-format="electronic" date-type="pub">
        <day>29</day>
        <month>09</month>
        <year>2005</year>
      </pub-date>
    </article-meta>
  </front>
  <body>
    <p>
      <bold>ARTIGO ORIGINAL</bold>
    </p>
    <p>Indicadores antropom&#233;tricos de obesidade como instrumento de triagem para risco coronariano elevado em adultos na cidade de Salvador &#150; Bahia </p>
    <p>
      <bold>Francisco Jos&#233; Gondim Pitanga; Ines Lessa </bold>
    </p>
    <p>Universidade Federal da Bahia - Salvador, BA</p>
    <p>
      <xref ref-type="corresp" rid="end">Endere&#231;o para correspond&#234;ncia</xref>
    </p>
    <p>
      <bold>RESUMO</bold>
    </p>
    <p><bold>OBJETIVO: </bold>Comparar v&#225;rios indicadores antropom&#233;tricos de obesidade e identificar dentre eles qual melhor discrimina o risco coronariano elevado (RCE). </p>
    <p><bold>M&#201;TODOS:</bold> Estudo de corte transversal, com amostra composta por 968 adultos de 30 a 74 anos de idade, sendo 391 (40,4%) do sexo masculino. Foram constru&#237;das diversas curvas <italic>Receiver Operating Characteristic</italic> (ROC) e comparadas &#224;s &#225;reas sob as mesmas entre o &#237;ndice de conicidade (&#237;ndice C), &#237;ndice de massa corporal (IMC), raz&#227;o circunfer&#234;ncia cintura-quadril (RCCQ), circunfer&#234;ncia de cintura (CC) e RCE. Verificou-se tamb&#233;m a sensibilidade e especificidade para identificar e comparar o melhor ponto de corte entre os diversos indicadores de obesidade para discriminar o RCE. Foi utilizado intervalo de confian&#231;a a 95%. </p>
    <p><bold>RESULTADOS:</bold> A maior &#225;rea sob a curva ROC foi encontrada entre o &#237;ndice C e RCE, em indiv&#237;duos do sexo masculino, 0,80 (0,74-0,85), diferindo significativamente dos demais indicadores de obesidade. Em mulheres, a maior &#225;rea sob a curva ROC encontrada foi de 0,76 (0,71-0,81), sendo iguais entre &#237;ndice C, RCCQ e RCE. </p>
    <p><bold>CONCLUS&#195;O:</bold> Esses resultados demonstram que o &#237;ndice C e RCCQ s&#227;o os melhores indicadores de obesidade para discriminar RCE. A CC tem intermedi&#225;rio poder discriminat&#243;rio e o IMC foi o indicador antropom&#233;trico de obesidade menos adequado para discriminar RCE. Estes dados sugerem que os indicadores de obesidade abdominal s&#227;o melhores para discriminar RCE que os indicadores de obesidade generalizada. </p>
    <p><bold>Palavras-chave: </bold>indicadores antropom&#233;tricos de obesidade, risco coronariano, obesidade abdominal</p>
    <p> O papel da obesidade como fator de risco coronariano &#233; controverso, no entanto, a melhor explica&#231;&#227;o para a associa&#231;&#227;o entre obesidade e doen&#231;a card&#237;aca isqu&#234;mica &#233; que esta ocorreria em um subgrupo de obesos, ou seja, nas pessoas que apresentassem adiposidade localizada na regi&#227;o abdominal ou central, mesmo na aus&#234;ncia da obesidade generalizada<sup>1</sup>. </p>
    <p>O ac&#250;mulo de gordura na regi&#227;o do abdome vem sendo descrito como o tipo de obesidade que oferece maior risco para a sa&#250;de dos indiv&#237;duos. A incid&#234;ncia de diabetes, aterosclerose, gota, c&#225;lculo urin&#225;rio e morte card&#237;aca s&#250;bita &#233; elevada em pessoas obesas, por&#233;m, um aspecto da adiposidade que desperta a aten&#231;&#227;o &#233; a distribui&#231;&#227;o regional da gordura no corpo. Considerando essa perspectiva, relacionou-se a espessura das dobras cut&#226;neas pesco&#231;o/regi&#227;o sacra e a propor&#231;&#227;o gordura-m&#250;sculos da regi&#227;o br&#225;quio/femural, sendo criado o &#237;ndice de diferencia&#231;&#227;o masculina. Atrav&#233;s desse &#237;ndice, classificou-se a gordura em (a) andr&#243;ide; aquela localizada na regi&#227;o central e mais espec&#237;fica para homens, e (b) gin&#243;ide; gordura mais localizada nos quadris e coxas, e mais espec&#237;fica para mulheres<sup>2</sup>. </p>
    <p>Alguns anos ap&#243;s<sup>3</sup>, a obesidade abdominal foi relacionada com o aumento do risco de infarto do mioc&#225;rdio, acidente vascular cerebral e morte prematura, encontrando-se forte associa&#231;&#227;o entre essas vari&#225;veis. Essas descobertas foram particularmente importantes na medida que, at&#233; ent&#227;o, existia associa&#231;&#227;o somente entre &#237;ndices de obesidade geral e as doen&#231;as mencionadas. Contrariando os conceitos da &#233;poca, o mais alto risco de infarto do mioc&#225;rdio ou morte prematura foi encontrado em homens com alta quantidade de gordura abdominal, sugerindo que homens, mesmo magros, por&#233;m, com gordura corporal concentrada no abdome s&#227;o aqueles com mais alto risco para desenvolver doen&#231;as cardiovasculares. </p>
    <p>Pesquisas mais recentes continuam a identificar a gordura abdominal como potente fator de risco coronariano, com a compara&#231;&#227;o entre os diversos indicadores de obesidade como fator de risco cardiovascular sendo fruto de numerosos estudos<sup>4-6</sup>. </p>
    <p>Nesse estudo, o principal objetivo foi analisar os indicadores antropom&#233;tricos de obesidade generalizada e abdominal como instrumentos de triagem para discriminar risco coronariano elevado (RCE). </p>
    <p>
      <bold>M&#233;todos</bold>
    </p>
    <p>Estudo de corte transversal realizado em Salvador-Bahia, Brasil, em 2000, em subgrupo de participantes do projeto "Monit", desenvolvido pela equipe de doen&#231;as cr&#244;nicas n&#227;o transmiss&#237;veis do ISC-UFBA e financiado pelo Minist&#233;rio da Sa&#250;de do Brasil, que teve como objetivo determinar a preval&#234;ncia dos fatores de risco cardiovasculares na cidade de Salvador<sup>7</sup>. </p>
    <p>Amostra probabil&#237;stica por conglomerados de 37 setores censit&#225;rios do per&#237;metro urbano da cidade de Salvador. Inicialmente, foram realizadas 2.476 entrevistas, havendo 72 recusas (2,9%) e extravio de 107 (4,7%) question&#225;rios com protocolo completo. A amostra no projeto original ficou constitu&#237;da por 2.297 adultos, com idades entre 20 e 74 anos de idade, sendo 1.025 do sexo masculino e 1.272 do sexo feminino. Como o indicador de risco coronariano utilizado no presente estudo foi constru&#237;do com base em popula&#231;&#227;o de 30 a 74 anos, a amostra do projeto original foi reduzida para 1.654 adultos, sendo 711 homens e 943 mulheres. Para o presente estudo foi utilizado um subgrupo deste grupo et&#225;rio, constitu&#237;do por pessoas que realizaram todo o protocolo de medidas proposto, totalizando 391 homens e 577 mulheres, com idade entre 30 e 74 anos, o que equivale a 55% de participantes do sexo masculino e 61% do sexo feminino, em rela&#231;&#227;o ao banco de dados original das faixas et&#225;rias em contexto. </p>
    <p>Todos os participantes da pesquisa foram entrevistados em domic&#237;lio para coleta dos dados demogr&#225;ficos e tiveram a press&#227;o arterial sist&#243;lica (PAS) e press&#227;o arterial diast&#243;lica (PAD) medidas por 6 (seis) vezes; as primeiras, tr&#234;s consecutivamente, 30 minutos ap&#243;s in&#237;cio da entrevista e as 3 seguintes ap&#243;s intervalo de 20 minutos, bra&#231;o esquerdo, sentados, bexiga vazia, sem terem fumado, tomado caf&#233; ou &#225;lcool nos 30 minutos precedentes &#224;s tomadas. A cintura foi medida estando o sujeito com o m&#237;nimo de roupa poss&#237;vel, na dist&#226;ncia m&#233;dia entre a &#250;ltima costela flutuante e a crista il&#237;aca, enquanto que o quadril foi medido com a fita m&#233;trica passando pelos troc&#226;nteres femurais (duas medidas) de cada circunfer&#234;ncia. Peso e estatura foram medidos no Centro de Sa&#250;de de cada bairro pela pr&#243;pria equipe do estudo (duas medidas). O &#237;ndice de conicidade (&#237;ndice C) foi determinado atrav&#233;s das medidas de peso, estatura e circunfer&#234;ncia da cintura (cc), utilizando-se a equa&#231;&#227;o matem&#225;tica<sup>8</sup>: </p>
    <disp-quote>
      <p>
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-e01.jpg"/></p>
    </disp-quote>
    <p>O &#237;ndice de massa corporal (IMC) foi determinado pela divis&#227;o do peso/estatura. A raz&#227;o circunfer&#234;ncia cintura-quadril (RCCQ) foi determinada pela divis&#227;o da circunfer&#234;ncia da cintura pela circunfer&#234;ncia do quadril. A coleta do sangue para os exames bioqu&#237;micos, ap&#243;s 12 horas de jejum, foi realizada no Centro de Sa&#250;de de suporte ao projeto. Foram realizadas dosagens de colesterol total (m&#233;todo Trinder enzim&#225;tico), HDL-C (m&#233;todo Labtest) e glicemia (m&#233;todo Trinder enzim&#225;tico). As t&#233;cnicas e m&#233;todos para os exames bioqu&#237;micos foram padronizados pela Sociedade Brasileira de Patologia Cl&#237;nica. O projeto foi apresentado ao Comit&#234; de &#201;tica do Conselho Regional de Medicina do Estado da Bahia, sendo aprovado na &#237;ntegra. Todos os participantes do estudo ou seus respons&#225;veis assinaram termo de consentimento concordando em participar da pesquisa. </p>
    <p>Para medida da tens&#227;o arterial foi utilizado tensi&#244;metro eletr&#244;nico (OMRON, modelo HEM-705CP). A estatura foi medida com estadi&#244;metro ingl&#234;s, tipo "Leicesters", acoplado a uma base para medida de estatura em campo, fabricado pela Child Growth Foundation. O peso corporal foi medido em balan&#231;as padr&#227;o, tipo "banheiro", novas e aferidas no in&#237;cio e periodicamente pelo INMETRO, com certificado pr&#243;prio, referindo margem de erro de&#177;100 g. Para medida de circunfer&#234;ncia de cintura e quadris foi utilizada fita m&#233;trica met&#225;lica e flex&#237;vel, marca Starrett, com defini&#231;&#227;o de medida de 0,1 cm. </p>
    <p>Foram utilizadas as seguintes vari&#225;veis: 1) idade; 2) sexo; 3) estatura; 4) peso; 5) CC; 6) circunfer&#234;ncia do quadril; 7) tabagismo; 8) glicemia; 9) colesterol total; 10) HDL-C; 11) PAS e 12) PAD. </p>
    <p>Utilizando algumas destas vari&#225;veis, foi constru&#237;do um indicador de risco coronariano para representar em conjunto os fatores de risco cardiovascular analisados simultaneamente. </p>
    <p> Para constru&#231;&#227;o do indicador de risco coronariano inicialmente foi utilizado algoritmo elaborado com base na coorte de Framinghan<sup>9</sup>, cujos autores analisaram dados prospectivos resultantes do acompanhamento, durante 12 anos, de 2.489 homens e 2.856 mulheres com idade de 30 a 74 anos. Com base em casos incidentes de doen&#231;a arterial coronariana (DAC), usaram o modelo de regress&#227;o de Cox e elaboraram tabela de pontua&#231;&#227;o (algoritmo) com a inclus&#227;o da idade, PAS, PAD, colesterol total, HDL-C, tabagismo e diabetes. A pontua&#231;&#227;o para cada vari&#225;vel pode ser positiva, fator de risco, ou negativa, fator de prote&#231;&#227;o. </p>
    <p>Na determina&#231;&#227;o dos pontos para a casu&#237;stica de Salvador, foi utilizado o coeficiente beta dos modelos de an&#225;lises de regress&#227;o de Cox<sup>9</sup>, com base nas pontua&#231;&#245;es das pr&#243;prias vari&#225;veis da amostra local. O percentil 80 (80% dos participantes da amostra apresentando pontua&#231;&#227;o abaixo do valor utilizado para identificar o alto risco coronariano) foi escolhido como ponto de corte para dicotomiza&#231;&#227;o da amostra em "risco coronariano elevado/n&#227;o elevado". Para o sexo masculino o percentil 80 correspondeu a 8 pontos e para o feminino a 10 pontos. </p>
    <p>Os pontos de corte para o &#237;ndice C, RCCQ, CC e IMC foram determinados atrav&#233;s das curvas Receiver Operating Characteristic (ROC), freq&#252;entemente utilizadas para determina&#231;&#227;o de pontos de corte em testes diagn&#243;sticos ou de triagem<sup>10</sup>. </p>
    <p>Inicialmente, foi identificada a &#225;rea total sob a curva ROC entre o &#237;ndice C, RCCQ, CC, IMC e RCE. Quanto maior &#225;rea sob a curva ROC, maior o poder discriminat&#243;rio dos indicadores de obesidade para RCE. Utilizou-se intervalo de confian&#231;a (IC) a 95%. O c&#225;lculo do IC a 95% determina se a capacidade preditiva dos indicadores de obesidade n&#227;o &#233; devida ao acaso, n&#227;o devendo o seu limite inferior ser menor do que 0,50<sup>11</sup>. </p>
    <p>Na seq&#252;&#234;ncia, foram calculadas a sensibilidade e especificidade, al&#233;m dos pontos de corte para o &#237;ndice C, RCCQ, CC, IMC e RCE. Valores identificados por interm&#233;dio da curva ROC constituem-se em pontos de corte que dever&#227;o promover um equil&#237;brio mais adequado entre sensibilidade e especificidade para o &#237;ndice C, RCCQ, CC e IMC como discriminadores de RCE. </p>
    <p>Os dados foram analisados atrav&#233;s do programa estat&#237;stico "STATA", vers&#227;o 7.0. </p>
    <p>
      <bold>Resultados</bold>
    </p>
    <p>As caracter&#237;sticas da amostra est&#227;o demonstradas na <xref ref-type="table" rid="tab01">tabela I</xref>. Observa-se que apesar de n&#227;o existirem diferen&#231;as estatisticamente significativas entre homens e mulheres nas vari&#225;veis escolaridade, ra&#231;a e risco coronariano, a maioria dos participantes da amostra apresenta baixo risco coronariano e n&#227;o s&#227;o brancos. Observam-se diferen&#231;as estatisticamente significativas entre homens e mulheres nas vari&#225;veis: peso corporal, estatura, circunfer&#234;ncia da cintura e do quadril, PAS, PAD, colesterol total, HDL-C, &#237;ndice C, IMC, RCCQ e CC. </p>
    <table-wrap id="tab01">
      <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt01.jpg"/></table-wrap>
    <p>As &#225;reas sob a curva ROC entre o &#237;ndice C, IMC, RCCQ, CC e RCE, em pessoas do sexo masculino, podem ser observadas na <xref ref-type="fig" rid="fig01">figura 1</xref>, notando-se que o &#237;ndice C apresenta o maior percentual sob a curva dentre os indicadores de obesidade para discriminar o RCE em homens, com diferen&#231;as estatisticamente significativas entre os mesmos (<xref ref-type="fig" rid="fig01">fig. 1</xref> e <xref ref-type="table" rid="tab02">tab. II</xref>). </p>
    <fig id="fig01">
      <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gf01.jpg"/></fig>
    <table-wrap id="tab02">
      <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt02.jpg"/></table-wrap>
    <p>Na <xref ref-type="fig" rid="fig02">figura 2</xref>, pode-se observar que para as mulheres as &#225;reas sob a curva ROC entre o &#237;ndice C, IMC, RCCQ, CC e o RCE s&#227;o semelhantes para &#237;ndice C e RCCQ, com diferen&#231;as estatisticamente significativas entre os diversos indicadores de obesidade para discriminar RCE (<xref ref-type="fig" rid="fig02">fig. 2</xref> e <xref ref-type="table" rid="tab03">tab. III</xref>). </p>
    <fig id="fig02">
      <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gf02.jpg"/></fig>
    <table-wrap id="tab03">
      <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt03.jpg"/></table-wrap>
    <p>Os valores de sensibilidade e especificidade com mais adequado equil&#237;brio entre si, al&#233;m dos seus respectivos pontos de cortes, s&#227;o apresentados para todos os indicadores de obesidade como discriminadores de RCE. Nota-se que, em rela&#231;&#227;o ao sexo masculino, o &#237;ndice C apresenta-se como o indicador de obesidade com melhores valores de sensibilidade e especificidade para discriminar RCE. Entre as mulheres, a RCCQ apresenta valores discretamente maiores do que o &#237;ndice C (<xref ref-type="table" rid="tab04">tab. IV</xref>). </p>
    <table-wrap id="tab04">
      <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt04.jpg"/></table-wrap>
    <p>
      <bold>Discuss&#227;o</bold>
    </p>
    <p>Com rela&#231;&#227;o aos pontos de corte para CC identificados no presente estudo, os resultados apresentam diverg&#234;ncias com a literatura. Existe proposta para valores de CC acima de 102 cm (homens) e 88 cm (mulheres) como melhores pontos de corte para identificar obesidade e RCE<sup>12</sup>. Na amostra analisada, a utiliza&#231;&#227;o destes valores implicaria uma grande quantidade de pessoas classificadas de forma incorreta, gerando muitos falsos-negativos, j&#225; que os pontos de corte identificados nos participantes do estudo foram de 88 cm para homens e 83 cm para mulheres. </p>
    <p>Ao ser feita a compara&#231;&#227;o entre os diversos indicadores de obesidade com RCE, observou-se que o &#237;ndice C e RCCQ s&#227;o os seus melhores discriminadores, apesar dos valores de sensibilidade e especificidade encontrados n&#227;o serem ideais. Existem alguns estudos comparativos entre os diversos indicadores de obesidade como discriminadores de RCE, por&#233;m, estudos que incluam o &#237;ndice C neste tipo de an&#225;lise n&#227;o foram localizados na literatura consultada. </p>
    <p>Em estudo realizado com 2.728 japoneses de ambos os sexos, os melhores pontos de corte observados pelos autores para IMC, CC e RCCQ, como discriminadores de risco cardiovascular elevado entre homens, foram: 23,5 kg/m&#178;, 84 cm e 0,90, enquanto que, entre as mulheres, os melhores pontos de corte foram: 22,5 kg/m, 72 cm e 0,80, respectivamente. Os autores conclu&#237;ram que os indicadores de adiposidade central discriminam melhor o RCE que os indicadores de obesidade total<sup>13</sup>. </p>
    <p>Nossos resultados apresentam algumas semelhan&#231;as com o estudo citado anteriormente, principalmente nos pontos de corte para IMC e RCCQ, entre homens, e RCCQ, entre mulheres. O ponto de corte observado para CC como discriminador de RCE entre mulheres do nosso estudo foi muito maior que o apresentado em japoneses. </p>
    <p>Em estudo similar<sup>14</sup>, realizado com aproximadamente 80.000 chineses, foram identificados os melhores pontos de corte para IMC e CC como discriminadores de RCE. Para o IMC, identificou-se o valor de 24 kg/m&#178; para ambos os sexos, enquanto que para CC identificou-se o valor de 85 cm para homens e 80 cm para mulheres, valores esses muito pr&#243;ximos aos encontrados em Salvador - Brasil. </p>
    <p>O poder discriminat&#243;rio do IMC e CC, como ferramentas de triagem para risco cardiovascular, foi analisado em 5.149 mulheres francesas<sup>15</sup>, encontrando-se &#225;reas sob a curva ROC similares para ambos os indicadores de obesidade, quando utilizados para discriminar mulheres com pelo menos um fator de risco cardiovascular: 0,71 (0,69-0,73) e 0,76 (0,74-0,78) para IMC e CC, respectivamente. Apesar das &#225;reas sob a curva ROC apresentarem valores aproximados, os autores relataram que a CC tem maior poder discriminat&#243;rio para classificar RCE que o IMC. Nossos resultados demonstram valores menores nas &#225;reas sob a curva ROC, provavelmente em fun&#231;&#227;o do agrupamento dos diversos fatores de risco cardiovascular em um &#250;nico indicador de risco coronariano. Os mesmos autores encontraram diferentes &#225;reas sob a curva ROC para cada um dos fatores de risco cardiovascular quando analisados isoladamente. </p>
    <p>Em estudo, realizado em Taiwan<sup>16</sup>,para determinar os melhores pontos de corte de diversas vari&#225;veis antropom&#233;tricas como discriminadores de risco cardiovascular, foram analisados 26.359 homens e 29.204 mulheres, encontrando-se os seguintes pontos de corte para discriminar pelo menos um fator de risco cardiovascular: IMC (23,6 e 22,1 kg/m), CC (80,5 e 71,5 cm) e RCCQ (0,85 e 0,76) para homens e mulheres, respectivamente. </p>
    <p>Os valores dos nossos pontos de corte foram mais elevados, demonstrando que na popula&#231;&#227;o da Taiwan o comportamento das vari&#225;veis antropom&#233;tricas como discriminadores de RCE apresenta caracter&#237;sticas diferentes da nossa popula&#231;&#227;o. </p>
    <p>Outros estudos foram realizados na perspectiva de identificar pontos de corte e comparar os diversos indicadores de obesidade como discriminadores de fatores de risco cardiovascular analisados individualmente. Na &#205;ndia<sup>17</sup>, identificou-se como pontos de corte para discriminar diabetes os valores de 85 cm para homens e 80 cm para mulheres, para CC, e de 0,88 (homens) e 0,81 (mulheres) para RCCQ. </p>
    <p>Em outro trabalho, com objetivo de comparar o &#237;ndice C e IMC como discriminador de hiperglicemia, em 1.325 adultos na cidade de Salvador-BA, encontrou-se o &#237;ndice C com maior poder discriminat&#243;rio do que o IMC, com &#225;reas sob a curva ROC de 0,71 (0,66-0,77) e 0,62 (0,57-0,67), respectivamente<sup>18</sup>. </p>
    <p>Em outro estudo<sup>19</sup>, usando a mesma popula&#231;&#227;o, comparou-se o &#237;ndice C e IMC como discriminador de hipertens&#227;o arterial sist&#243;lica, encontrando-se &#225;reas sob a curva ROC sem diferen&#231;as estatisticamente significativas, demonstrando que o poder discriminat&#243;rio para RCE entre &#237;ndice C e IMC &#233; similar. </p>
    <p>Uma das limita&#231;&#245;es do presente estudo foi &#224; classifica&#231;&#227;o do RCE usando o algoritmo proposto com base na coorte de Framinghan<sup>9</sup>. Utilizamos este algoritmo aplicando seus valores nos indiv&#237;duos participantes da nossa amostra e definimos que o RCE seria identificado por meio da pontua&#231;&#227;o referente ao percentil 80 na distribui&#231;&#227;o de percentis da pr&#243;pria popula&#231;&#227;o de estudo. Existem controv&#233;rsias quanto &#224; pontua&#231;&#227;o ideal para identifica&#231;&#227;o do RCE em curto prazo. Existem tamb&#233;m sugest&#245;es<sup>20</sup> de que o RCE poderia ser identificado atrav&#233;s da pontua&#231;&#227;o equivalente a 20% do risco absoluto para coronariopatia nos pr&#243;ximos 10 anos, o que equivaleria a 9 pontos para homens e 15 pontos para mulheres, no algoritmo proposto com base na coorte de Framinghan<sup>9</sup>. No nosso estudo, o RCE foi identificado atrav&#233;s do ponto de corte de 8 pontos para homens e 10 pontos para mulheres. </p>
    <p>Estudos recentes demonstram que a predi&#231;&#227;o do RCE, em longo prazo, feita atrav&#233;s do algoritmo proposto com base na coorte de Framinghan<sup>9</sup>, pode ser generalizada para outras popula&#231;&#245;es, bem como pode ser utilizada em estudos transversais. A generaliza&#231;&#227;o para outras popula&#231;&#245;es deve ser feita ap&#243;s recalibra&#231;&#227;o para diferentes preval&#234;ncias dos fatores de risco envolvidos na an&#225;lise<sup>21, 22</sup>. </p>
    <p>Com rela&#231;&#227;o &#224;s vari&#225;veis utilizadas no algoritmo proposto, o ponto de corte adotado para diabetes foi o valor de 140 mg/dl, enquanto neste estudo o ponto de corte adotado para diabetes foi o valor de 126 mg/dl, considerando as recentes sugest&#245;es de diversas associa&#231;&#245;es cient&#237;ficas<sup>19, 23, 24</sup>. </p>
    <p>Diversas outras vari&#225;veis podem ser inclu&#237;das em modelos preditivos de risco coronariano. Hist&#243;ria familiar prematura de DAC, apesar de ser fortemente associada a eventos coronarianos, n&#227;o est&#225; dispon&#237;vel nos sujeitos analisados no algoritmo proposto com base na coorte de Framinghan<sup>8</sup>. O fibrinog&#234;nio tem sido associado ao risco cardiovascular<sup>23</sup>, por&#233;m, tamb&#233;m n&#227;o est&#225; dispon&#237;vel no estudo citado anteriormente<sup>8</sup>. Al&#233;m disso, devido a sua alta variabilidade biol&#243;gica e aus&#234;ncia de comprova&#231;&#227;o de que interven&#231;&#245;es espec&#237;ficas diminuam o risco cardiovascular, n&#227;o se recomenda sua determina&#231;&#227;o rotineira<sup>23</sup>. Hipertrofia do ventr&#237;culo esquerdo, observada por meio do ECG, tem sido usada em algoritmos para predi&#231;&#227;o de DAC, por&#233;m, considerando sua alta associa&#231;&#227;o com hipertens&#227;o arterial, n&#227;o foi utilizada no algoritmo citado anteriormente<sup>8</sup>. </p>
    <p>Por outro lado, ao ser feita compara&#231;&#227;o de diferentes m&#233;todos de an&#225;lise de RCE, sugere-se a incorpora&#231;&#227;o da raz&#227;o colesterol total/HDL-C nos algoritmos para classifica&#231;&#227;o de RCE<sup>25</sup>. No nosso estudo, utilizou-se a pontua&#231;&#227;o do colesterol total separadamente do HDL-C. </p>
    <p>De acordo com os resultados encontrados ficam evid&#234;ncias de que os indicadores de obesidade central, principalmente &#237;ndice C e RCCQ, s&#227;o melhores para discriminar RCE que indicadores de obesidade generalizada (IMC). A CC tem desempenho razo&#225;vel, apesar de diversos estudos terem demonstrado que a mesma &#233; um potente discriminador de RCE (dados n&#227;o confirmados em nossas observa&#231;&#245;es). Estes resultados sugerem que os indicadores de obesidade abdominal s&#227;o melhores para discriminar RCE que os indicadores de obesidade generalizada. Sugerem-se novos estudos com crian&#231;as, adolescentes e adultos de diferentes grupos et&#225;rios para identificar e comparar o poder discriminat&#243;rio dos indicadores antropom&#233;tricos de obesidade e RCE nestes subgrupos espec&#237;ficos. </p>
    <p>
      <fn id="end" fn-type="corresp">
        <label>
          <bold>Endere&#231;o para correspond&#234;ncia</bold>
        </label>
        <p> Francisco Jos&#233; Gondim Pitanga </p>
        <p> Av. Luiz Tarqu&#237;nio Pontes, 600 </p>
        <p> 42700-000 - Lauro de Freitas, BA </p>
        <p> E-mail: </p>
        <p>
          <email xlink:href="pitanga@lognet.com.br">pitanga@lognet.com.br</email>
        </p>
      </fn>
    </p>
    <p>Recebido em 11/05/2004 </p>
    <p> Aceito em 21/01/2005</p>
  </body>
  <back>
    <ref-list>
      <ref id="B1">
        <mixed-citation>1. Lessa I. O Adulto brasileiro e as doen&#231;as    da modernidade: epidemiologia das doen&#231;as cr&#244;nicas n&#227;o-transmiss&#237;veis.    S&#227;o Paulo: Editora Hucitec-Abrasco; 1998. </mixed-citation>
        <element-citation publication-type="book">
          <source>O Adulto brasileiro e as doen&#231;as da modernidade: epidemiologia das doen&#231;as cr&#244;nicas n&#227;o-transmiss&#237;veis</source>
          <date>
            <year>1998</year>
          </date>
          <person-group person-group-type="author">
            <name>
              <surname>Lessa</surname>
              <given-names>I</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B2">
        <mixed-citation>2. Vague J. The degree of masculine differentiation    of obesities: A factor determining predisposition to diabetes, atherosclerosis,    gout, and uric calculous disease. Am J Clin Nutr 1956; 4: 20-34. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>The degree of masculine differentiation of obesities: A factor determining predisposition to diabetes, atherosclerosis, gout, and uric calculous disease</article-title>
          <source>Am J Clin Nutr</source>
          <date>
            <year>1956</year>
          </date>
          <fpage>20</fpage>
          <lpage>34</lpage>
          <volume>4</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Vague</surname>
              <given-names>J</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B3">
        <mixed-citation>3. Larsson B, Svardsudd K, Welin L, Wilhelmsen    L, Bjorntorp P, Tibblin G. Abdominal adipose tissue distribution, obesity, and    risk of cardiovascular disease and death: 13 year follow up of participants    in the study of men born in 1913. Br Med J 1984; 288: 1401-4. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13 year follow up of participants in the study of men born in 1913</article-title>
          <source>Br Med J</source>
          <date>
            <year>1984</year>
          </date>
          <fpage>1401</fpage>
          <lpage>4</lpage>
          <volume>288</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Larsson</surname>
              <given-names>B</given-names>
            </name>
            <name>
              <surname>Svardsudd</surname>
              <given-names>K</given-names>
            </name>
            <name>
              <surname>Welin</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Wilhelmsen</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Bjorntorp</surname>
              <given-names>P</given-names>
            </name>
            <name>
              <surname>Tibblin</surname>
              <given-names>G</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B4">
        <mixed-citation>4. Gustat J, Elkasabany A, Srinivasan S, Bereson    S. Relation of abdominal heigth to cardiovascular risk factors in young adults.    Am J Epidemiol 2000; 151: 885-91. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Relation of abdominal heigth to cardiovascular risk factors in young adults</article-title>
          <source>Am J Epidemiol</source>
          <date>
            <year>2000</year>
          </date>
          <fpage>885</fpage>
          <lpage>91</lpage>
          <volume>151</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Gustat</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Elkasabany</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Srinivasan</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Bereson</surname>
              <given-names>S</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B5">
        <mixed-citation>5. Visscher Tl, Seidell JC, Molarius A, Van Der    Kuip D, Hofman A, Witteman JC. A comparison of body mass index, waist-hip ratio    and waist circumference as predictors of all-cause mortality among the elderly:    The Rotterdam study. Int J Obes Relat Metab Disord 2001; 25: 1730-5. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>A comparison of body mass index, waist-hip ratio and waist circumference as predictors of all-cause mortality among the elderly: The Rotterdam study</article-title>
          <source>Int J Obes Relat Metab Disord</source>
          <date>
            <year>2001</year>
          </date>
          <fpage>1730</fpage>
          <lpage>5</lpage>
          <volume>25</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Visscher</surname>
              <given-names>Tl</given-names>
            </name>
            <name>
              <surname>Seidell</surname>
              <given-names>JC</given-names>
            </name>
            <name>
              <surname>Molarius</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Van Der Kuip</surname>
              <given-names>D</given-names>
            </name>
            <name>
              <surname>Hofman</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Witteman</surname>
              <given-names>JC</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B6">
        <mixed-citation>6. Poulter N. Global risk of cardiovascular disease.    Heart 2003; 89 (suppl 2): 112-5. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Global risk of cardiovascular disease</article-title>
          <source>Heart</source>
          <date>
            <year>2003</year>
          </date>
          <fpage>112</fpage>
          <lpage>5</lpage>
          <volume>89</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Poulter</surname>
              <given-names>N</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B7">
        <mixed-citation>7. Centro Nacional de Epidemiologia, Funda&#231;&#227;o    Nacional de Sa&#250;de. Projeto Monit: Relat&#243;rio de Pesquisa. Bras&#237;lia:    MS; Salvador: ISC/UFBA; 2000. </mixed-citation>
        <element-citation publication-type="book">
          <source>Projeto Monit: Relat&#243;rio de Pesquisa</source>
          <date>
            <year>2000</year>
          </date>
        </element-citation>
      </ref>
      <ref id="B8">
        <mixed-citation>8. Valdez R. A simple model-based index of abdominal    adiposity. J Clin Epidemiol 1991; 44: 955-6. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>A simple model-based index of abdominal adiposity</article-title>
          <source>J Clin Epidemiol</source>
          <date>
            <year>1991</year>
          </date>
          <fpage>955</fpage>
          <lpage>6</lpage>
          <volume>44</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Valdez</surname>
              <given-names>R</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B9">
        <mixed-citation>9. Wilson PWF, D'Agostino RB, Levy D, Belanger    AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk    factores categories. Circulation 1998; 97: 1837-47. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Prediction of coronary heart disease using risk factores categories</article-title>
          <source>Circulation</source>
          <date>
            <year>1998</year>
          </date>
          <fpage>1837</fpage>
          <lpage>47</lpage>
          <volume>97</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Wilson</surname>
              <given-names>PWF</given-names>
            </name>
            <name>
              <surname>D'Agostino</surname>
              <given-names>RB</given-names>
            </name>
            <name>
              <surname>Levy</surname>
              <given-names>D</given-names>
            </name>
            <name>
              <surname>Belanger</surname>
              <given-names>AM</given-names>
            </name>
            <name>
              <surname>Silbershatz</surname>
              <given-names>H</given-names>
            </name>
            <name>
              <surname>Kannel</surname>
              <given-names>WB</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B10">
        <mixed-citation>10. Erdreich LS, Lee ET. Use of relative operating    characteristics analysis in epidemiology: a method for dealing with subjective    judgement. Am J Epidemiol 1981; 114: 649-62. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Use of relative operating characteristics analysis in epidemiology: a method for dealing with subjective judgement</article-title>
          <source>Am J Epidemiol</source>
          <date>
            <year>1981</year>
          </date>
          <fpage>649</fpage>
          <lpage>62</lpage>
          <volume>114</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Erdreich</surname>
              <given-names>LS</given-names>
            </name>
            <name>
              <surname>Lee</surname>
              <given-names>ET</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B11">
        <mixed-citation>11. Schisterman EF, Faraggi D, Reiser B, Trevisan    M. Statistical inference for the area under the receiver operating characteristic    curve in the presence of random measurement error. Am J Epidemiol 2001; 154:    174-9. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Statistical inference for the area under the receiver operating characteristic curve in the presence of random measurement error</article-title>
          <source>Am J Epidemiol</source>
          <date>
            <year>2001</year>
          </date>
          <fpage>174</fpage>
          <lpage>9</lpage>
          <volume>154</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Schisterman</surname>
              <given-names>EF</given-names>
            </name>
            <name>
              <surname>Faraggi</surname>
              <given-names>D</given-names>
            </name>
            <name>
              <surname>Reiser</surname>
              <given-names>B</given-names>
            </name>
            <name>
              <surname>Trevisan</surname>
              <given-names>M</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B12">
        <label>12</label>
        <mixed-citation>12. NHLBI Obesity Education Initiative Expert Panel. Clinical Guidelines on Identification, Evaluation and Treatment of Overweight and Obesity in Adults. The Evidence Report. Bethesda, Md: National Institute of Health, National Heart, Lung and Blood Institute; 1998.</mixed-citation>
        <element-citation publication-type="book">
          <source>Clinical Guidelines on Identification, Evaluation and Treatment of Overweight and Obesity in Adults: The Evidence Report</source>
          <date>
            <year>1998</year>
          </date>
        </element-citation>
      </ref>
      <ref id="B13">
        <mixed-citation>13. Ito H, Nakasuga K, Ohshima A et al. Detection    of cardiovascular risk factors by indices of obesity obtained from anthropometry    and dual-energy x-ray absorptiometry in Japanese individuals. Int J Obes Relat    Metab Disord 2003; 27: 232-7. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Detection of cardiovascular risk factors by indices of obesity obtained from anthropometry and dual-energy x-ray absorptiometry in Japanese individuals</article-title>
          <source>Int J Obes Relat Metab Disord</source>
          <date>
            <year>2003</year>
          </date>
          <fpage>232</fpage>
          <lpage>7</lpage>
          <volume>27</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Ito</surname>
              <given-names>H</given-names>
            </name>
            <name>
              <surname>Nakasuga</surname>
              <given-names>K</given-names>
            </name>
            <name>
              <surname>Ohshima</surname>
              <given-names>A</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B14">
        <mixed-citation>14. Bei-Fan Z. Predictives values of body mass    index and waist circumference for risk factors of certain related disease in    Chinese adults: study on optimal cut-off points of body mass index and waist    circumference in Chinese adults. Asia Pac J Clin Nutr 2002; 11(suppl 8): S685-93.    </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Predictives values of body mass index and waist circumference for risk factors of certain related disease in Chinese adults: study on optimal cut-off points of body mass index and waist circumference in Chinese adults</article-title>
          <source>Asia Pac J Clin Nutr</source>
          <date>
            <year>2002</year>
          </date>
          <fpage>S685</fpage>
          <lpage>93</lpage>
          <issue>suppl 8</issue>
          <volume>11</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Bei-Fan</surname>
              <given-names>Z</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B15">
        <mixed-citation>15. Foucan L, Hanley J, Deloumeaux J, Suissa    S. Body mass index (BMI) and waist circumference (WC) as screening tools for    cardiovascular risk factors in Guadeloupean women. J Clin Epidemiol 2002; 55:    990-6. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Body mass index (BMI) and waist circumference (WC) as screening tools for cardiovascular risk factors in Guadeloupean women</article-title>
          <source>J Clin Epidemiol</source>
          <date>
            <year>2002</year>
          </date>
          <fpage>990</fpage>
          <lpage>6</lpage>
          <volume>55</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Foucan</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Hanley</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Deloumeaux</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Suissa</surname>
              <given-names>S</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B16">
        <mixed-citation>16. Lin WY, Lee LT, Chen CI et al. Optimal cut-off    values for obesity: using simple anthropometric indices to predict cardiovascular    risk factors in Taiwan. Int J Obes Relat Metab Disord 2002; 26: 1232-38. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Optimal cut-off values for obesity: using simple anthropometric indices to predict cardiovascular risk factors in Taiwan</article-title>
          <source>Int J Obes Relat Metab Disord</source>
          <date>
            <year>2002</year>
          </date>
          <fpage>1232</fpage>
          <lpage>38</lpage>
          <volume>26</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Lin</surname>
              <given-names>WY</given-names>
            </name>
            <name>
              <surname>Lee</surname>
              <given-names>LT</given-names>
            </name>
            <name>
              <surname>Chen</surname>
              <given-names>CI</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B17">
        <mixed-citation>17. Snehalatha C, Viswanathan V, Ramachadran    A. Cutoff values for normal anthropometric variables in Asian Indian adults.    Diabetes Care 2003; 26: 1380-4. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Cutoff values for normal anthropometric variables in Asian Indian adults</article-title>
          <source>Diabetes Care</source>
          <date>
            <year>2003</year>
          </date>
          <fpage>1380</fpage>
          <lpage>4</lpage>
          <volume>26</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Snehalatha</surname>
              <given-names>C</given-names>
            </name>
            <name>
              <surname>Viswanathan</surname>
              <given-names>V</given-names>
            </name>
            <name>
              <surname>Ramachadran</surname>
              <given-names>A</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B18">
        <mixed-citation>18. Pitanga FJG, Lessa I. An&#225;lise da Sensibilidade    e Especificidade entre &#205;ndice de Conicidade, &#205;ndice de Massa Corporal    e Hiperglicemia em Adultos de ambos os sexos. In: Resumos do XV Congresso de    Cardiologia do Estado da Bahia; 2003; Salvador: Sociedade Brasileira de Cardiologia;    2003a: p. 13. </mixed-citation>
        <element-citation publication-type="confproc">
          <date>
            <year>2003</year>
          </date>
          <fpage>13</fpage>
          <person-group person-group-type="author">
            <name>
              <surname>Pitanga</surname>
              <given-names>FJG</given-names>
            </name>
            <name>
              <surname>Lessa</surname>
              <given-names>I</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B19">
        <mixed-citation>19. Pitanga FJG, Lessa I. An&#225;lise da Sensibilidade    e Especificidade entre &#205;ndice de Conicidade, &#205;ndice de Massa Corporal    e Hipertens&#227;o Arterial Sist&#243;lica em Adultos de ambos os sexos.    In: Resumos do XV Congresso de Cardiologia do Estado da Bahia; 2003; Salvador:    Sociedade Brasileira de Cardiologia; 2003b: p. 14. </mixed-citation>
        <element-citation publication-type="confproc">
          <date>
            <year>2003</year>
          </date>
          <fpage>14</fpage>
          <person-group person-group-type="author">
            <name>
              <surname>Pitanga</surname>
              <given-names>FJG</given-names>
            </name>
            <name>
              <surname>Lessa</surname>
              <given-names>I</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B20">
        <mixed-citation>20. Grundy SM, Pasternak R, Greenland P, Smith    S, Fuster V. Assessment of cardiovascular risk by use of multiple-risk-factor    assessment equations. J Am Coll Cardiol 1999; 34: 1348-59. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Assessment of cardiovascular risk by use of multiple-risk-factor assessment equations</article-title>
          <source>J Am Coll Cardiol</source>
          <date>
            <year>1999</year>
          </date>
          <fpage>1348</fpage>
          <lpage>59</lpage>
          <volume>34</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Grundy</surname>
              <given-names>SM</given-names>
            </name>
            <name>
              <surname>Pasternak</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Greenland</surname>
              <given-names>P</given-names>
            </name>
            <name>
              <surname>Smith</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Fuster</surname>
              <given-names>V</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B21">
        <mixed-citation>21. D'Agostino RB, Grundy S, Sullivan LM, Wilson    P. Validation of the Framingham coronary Herat disease prediction scores. JAMA    2001; 286:180-7. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Validation of the Framingham coronary Herat disease prediction scores</article-title>
          <source>JAMA</source>
          <date>
            <year>2001</year>
          </date>
          <fpage>180</fpage>
          <lpage>7</lpage>
          <volume>286</volume>
          <person-group person-group-type="author">
            <name>
              <surname>D'Agostino</surname>
              <given-names>RB</given-names>
            </name>
            <name>
              <surname>Grundy</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Sullivan</surname>
              <given-names>LM</given-names>
            </name>
            <name>
              <surname>Wilson</surname>
              <given-names>P</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B22">
        <mixed-citation>22. Cappuccio FP, Oakeshott P, Kerry SM. Application    of Framingham risk estimates to ethnic minorities in United Kingdom and implications    for primary prevention of heart disease in general pratic: cross sectional population    based study. Br Med J 2002; 325: 1271-80. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Application of Framingham risk estimates to ethnic minorities in United Kingdom and implications for primary prevention of heart disease in general pratic: cross sectional population based study</article-title>
          <source>Br Med J</source>
          <date>
            <year>2002</year>
          </date>
          <fpage>1271</fpage>
          <lpage>80</lpage>
          <volume>325</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Cappuccio</surname>
              <given-names>FP</given-names>
            </name>
            <name>
              <surname>Oakeshott</surname>
              <given-names>P</given-names>
            </name>
            <name>
              <surname>Kerry</surname>
              <given-names>SM</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B23">
        <mixed-citation>23. American Diabetes Association. Report of    the expert committee on diagnosis and classification of diabetes mellitus. Diabetes    Care 1997; 20: 1183-202. </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Association: Report of the expert committee on diagnosis and classification of diabetes mellitus</article-title>
          <source>Diabetes Care</source>
          <date>
            <year>1997</year>
          </date>
          <fpage>1183</fpage>
          <lpage>202</lpage>
          <volume>20</volume>
          <person-group person-group-type="author">
            <name>
              <surname>American</surname>
              <given-names>Diabetes</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
      <ref id="B24">
        <mixed-citation>24. Sociedade Brasileira De Cardiologia. III    Diretrizes Brasileiras sobre dislipidemias e diretriz de preven&#231;&#227;o    da aterosclerose do Departamento de Aterosclerose. Arq Bras Cardiol 2001; 77(Suppl    3). </mixed-citation>
        <element-citation publication-type="journal">
          <article-title>III Diretrizes Brasileiras sobre dislipidemias e diretriz de preven&#231;&#227;o da aterosclerose do Departamento de Aterosclerose</article-title>
          <source>Arq Bras Cardiol</source>
          <date>
            <year>2001</year>
          </date>
          <volume>77</volume>
        </element-citation>
      </ref>
      <ref id="B25">
        <mixed-citation>25. Haq IU, Ransay LE, Jackson PR, Wallis EJ.    Prediction of coronary risk for primary prevention of coronary heart disease:    a comparison of methods. QJM 1999; 92: 379-85.</mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Prediction of coronary risk for primary prevention of coronary heart disease: a comparison of methods</article-title>
          <source>QJM</source>
          <date>
            <year>1999</year>
          </date>
          <fpage>379</fpage>
          <lpage>85</lpage>
          <volume>92</volume>
          <person-group person-group-type="author">
            <name>
              <surname>Haq</surname>
              <given-names>IU</given-names>
            </name>
            <name>
              <surname>Ransay</surname>
              <given-names>LE</given-names>
            </name>
            <name>
              <surname>Jackson</surname>
              <given-names>PR</given-names>
            </name>
            <name>
              <surname>Wallis</surname>
              <given-names>EJ</given-names>
            </name>
          </person-group>
        </element-citation>
      </ref>
    </ref-list>
  </back>
  <sub-article article-type="translation" id="TRen" xml:lang="en">
    <front-stub>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Original Articles</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title xml:lang="en">Anthropometric indexes of obesity as an instrument of screening for high coronary risk in adults in the city of Salvador - Bahia</article-title>
      </title-group>
      <abstract xml:lang="en">
        <p>OBJECTIVE: To compare some anthropometric indexes of obesity and identify among them which one best discriminates the high coronary risk (HCR). METHODS: A cross-section study, with sample consisting of 968 adults, between 30 and 74 years old, being 391 (40.4%) men. Many Receiver Operating Characteristic (ROC) curves were obtained and compared to areas under them among the conicity index (C index), body mass index (BMI), waist-hip circumference ratio (WHCR), waist circumference (WC) and HCR. The sensitivity and specificity to identify and compare the best cut-off point among the many indexes of obesity to discriminate the HCR were also identified. A confidence interval of 95% was used. RESULTS: The largest area under ROC curve was found between the C index and the HCR, in individuals of male sex, 0.80 (0.74-0.85), significantly differing from the other indexes of obesity. In women, the largest area found under the ROC curve was 0.76 (0.71-0.81), being equal between C, WHCR and HCR indexes. CONCLUSION: Those results show that C and WHCR indexes are the best indexes of obesity to discriminate HCR. WC has intermediate discriminatory power and the BMI was the least suitable anthropometric index of obesity to discriminate HCR. Those data suggest that the indexes of abdominal obesity are better to discriminate HCR than the indexes of general obesity.</p>
      </abstract>
      <kwd-group xml:lang="en">
        <kwd>anthropometric indexes of obesity</kwd>
        <kwd>coronary risk</kwd>
        <kwd>abdominal obesity</kwd>
      </kwd-group>
    </front-stub>
    <body>
      <p>
        <bold>ORIGINAL ARTICLE</bold>
      </p>
      <p>Anthropometric indexes of obesity as an instrument of screening for high coronary risk in adults in the city of Salvador &#150; Bahia</p>
      <p>
        <bold>Francisco Jos&#233; Gondim Pitanga; Ines Lessa </bold>
      </p>
      <p>Universidade Federal da Bahia - Salvador, BA - Brazil</p>
      <p>
        <xref ref-type="corresp" rid="end-body2">Correspondence</xref>
      </p>
      <p>
        <bold>ABSTRACT</bold>
      </p>
      <p><bold>OBJECTIVE: </bold>To compare some anthropometric indexes of obesity and identify among them which one best discriminates the high coronary risk (HCR). </p>
      <p><bold>METHODS: </bold>A cross-section study, with sample consisting of 968 adults, between 30 and 74 years old, being 391 (40.4%) men. Many Receiver Operating Characteristic (ROC) curves were obtained and compared to areas under them among the conicity index (C index), body mass index (BMI), waist-hip circumference ratio (WHCR), waist circumference (WC) and HCR. The sensitivity and specificity to identify and compare the best cut-off point among the many indexes of obesity to discriminate the HCR were also identified. A confidence interval of 95% was used. </p>
      <p><bold>RESULTS: </bold>The largest area under ROC curve was found between the C index and the HCR, in individuals of male sex, 0.80 (0.74-0.85), significantly differing from the other indexes of obesity. In women, the largest area found under the ROC curve was 0.76 (0.71-0.81), being equal between C, WHCR and HCR indexes. </p>
      <p><bold>CONCLUSION: </bold>Those results show that C and WHCR indexes are the best indexes of obesity to discriminate HCR. WC has intermediate discriminatory power and the BMI was the least suitable anthropometric index of obesity to discriminate HCR. Those data suggest that the indexes of abdominal obesity are better to discriminate HCR than the indexes of general obesity. </p>
      <p><bold>Key words: </bold>anthropometric indexes of obesity, coronary risk, abdominal obesity</p>
      <p>The role of obesity was coronary risk factor is controversial. However, the best explanation for the association between obesity and ischemic heart disease is that it would occur in a subgroup of obese individuals, which means in people who showed adiposity located in the abdominal or central region, even in the absence of general obesity<sup>1</sup>. </p>
      <p>The accumulation of fat in the abdomen region has been described as the type of obesity that offers the greatest risk for the health of the individuals. The incidence of diabetes, atherosclerosis, gout, urinary calculus and sudden cardiac death is high in some obese individuals. However, an adiposity aspect that draws attention to is the regional distribution of fat in the body. Considering that perspective, the thickness of neck/sacral region skin folds and the fat-muscle proportion of brachio/femoral region were related, which created the masculine differentiation index. Through that index, fat was classified as (a) android; that located in the central region and more specific for men, and (b) gynoid; fat located more in the hips and thighs, and more specific for women<sup>2</sup>. </p>
      <p>Some years later<sup>3</sup>, abdominal obesity was related to the increase of risk of myocardial infarction, cerebrovascular accident and early death, with a strong association among those variables. Such findings were particularly important to the extent that, by then, there was an association only between general obesity levels and the mentioned diseases. In opposition to the concepts at that time, the highest risk of myocardial infarction or early death was found in men with high quantity of abdominal fat, which suggested that men, even lean, but with body fat concentrated in the abdomen are those with the highest risk to develop cardiovascular diseases. </p>
      <p>More recent researches still identify the abdominal fat as a powerful coronary risk factor, with the comparison among the many obesity indexes as cardiovascular risk factor, being the result of many studies<sup>4,5,6</sup>. </p>
      <p>In this study, the main objective was to analyze the anthropometric indexes of general and abdominal obesity as instruments of screening to discriminate high coronary risk (HCR). </p>
      <p>
        <bold>Methods</bold>
      </p>
      <p>A cross-section study performed in Salvador, Bahia, Brazil, in 2000, in a subgroup of participants in "Monit" project, develop by the team of chronic non-transmitted diseases from ISC-UFBA and financed by the Brazilian Health Department, which had as objective to determine the prevalence of cardiovascular risk factors in the city of Salvador<sup>7</sup>. </p>
      <p> A probabilistic sample per conglomerates from 37 census sectors in the urban perimeter in the city of Salvador. Initially 2,476 interviews were carried out, with 72 refusals (2.9%) and loss of 107 (4.7%) questionnaires with complete protocol. The sample in the original project consisted of 2,297 adults, with age between 20 and 74 years old, with 1,025 of male sex and 1,272 of female sex. As the index of coronary risk used in the present study was constructed based on a population between 30 to 74 years old, the sample of the original project was reduced to 1,654 adults, 711 men and 943 women. For the present study, a subgroup from that age group was used, which consisted of people who were submitted to the whole measurement protocol proposed, amounting 391 men and 577 women, with age between 30 and 74 years old, which is equal to 55% of participants of male sex and 615 of female sex, in relation to the original database of the age ranges in context. </p>
      <p>All participants in the research were interviewed at home for the collection of demographic data and had the systolic blood pressure (SBP) and diastolic blood pressure (DBP) measured 6 times; the first three in a row, 30 minutes after the beginning of the interview and the 3 next ones after an interval of 20 minutes, left arm, seated, empty bladder, without having smoked, had a coffee or alcohol in the 30 minutes prior to the measurements. The waist was measured with the individual wearing the minimum clothes as possible, in the middle distance between the last floating rib and the iliac crest, whereas the hips were measured with the measuring tape passing on femoral trochanters (two measurements) of each circumference. Weight and height were measured in the Health Center of each neighborhood by the study team itself (two measurements). The conicity index (C index) was determined through the measurements of weight, height and waist circumference (CC), by using the mathematical equation<sup>8</sup>: </p>
      <disp-quote>
        <p>
          <graphic xlink:href="1678-4170-abc-85-01-0026-00006-e01-en.jpg"/></p>
      </disp-quote>
      <p>The body mass index (BMI) was determined through the division of weight/height. The waist-hip circumference ratio (WHCR) was determined through the division of the waist circumference by the hip circumference. The blood drawing for the biochemical tests, after 12 hours of feast, was carried out at Health Center for support to the project. Dosages of total cholesterol (enzymatic Trinder method), HDL-C (Labtest method) and glycemia (enzymatic Trinder method) were done. Techniques and methods for biochemical exams were standardized by Sociedade Brasileira de Patologia Cl&#237;nica &#150; Brazilian Association of Clinical Pathology. The project was submitted to the Comit&#234; de &#201;tica do Conselho Regional de Medicina do Estado da Bahia &#150; Ethics Committee of Medicine Regional Council of the State of Bahia, being fully approved. All participants in the study or their responsible signed a consent term agreeing on participating in the research. </p>
      <p>For the measurement of blood pressure an electronic tensiometer (OMRON, model HEM-705CP) was used. Height was measured with an English Leicester-type stadiometer, coupled to a base for field height measurement, manufactured by Child Growth Foundation. Body weight was measured in new bathroom-type, standard scales, gauged at the beginning and periodically by INMETRO, with their own certificate, referring to an error margin of&#177;100 g. For the measurement of waist and hip circumference, a metallic and flexible measurement tape, Starrett brand, with measurement definition of 0.1 cm. </p>
      <p>The following variables were used: 1) age; 2) sex; 3) height; 4) weight; 5) CC; 6) hip circumference; 7) smoking; 8) glycemia; 9) total cholesterol; 10) HDL-C; 11) SBP and 12) DBP. </p>
      <p>By using some of those variables, a coronary risk index was built to represent altogether the cardiovascular risk factors simultaneously analyzed. </p>
      <p>For the construction of coronary risk index initially an algorithm elaborated based on the cohort of Framinghan<sup>9</sup> was used, whose authors analyzed prospective data resulting from the follow-up, during 12 years, of 2,489 men and 2,856 women, with age of 30 to 74 years old. Based on incident cases of coronary artery disease (CAD), they used the regression model of Cox and elaborated a point table (algorithm) with the inclusion of age, SBP, DBP, total cholesterol, HDL-C, smoking and diabetes. The points for each variable can be positive, risk factor, or negative, protection factor. </p>
      <p>In determining the points for the casuistry of Salvador, the beta coefficient of the models of regression analysis of Cox<sup>9</sup> was used, based on the points of the variables themselves from the local sample. Percentile 80 (80% from the participants in the sample showing points lower than the value used to identify the high coronary risk) was chosen as cut-off point for dichotomization of the sample under "high/non-high coronary risk." For male sex the percentile 80 corresponded to 8 points and for the female, 10 points. </p>
      <p>Cut-off points for C index, WHCR, WC and BMI were determined through Receiver Operating Characteristic (ROC) curves, frequently used to determine cut-off points in diagnostic or screening tests<sup>10</sup>. </p>
      <p>Initially, the total area under the ROC curve between the C index, WHCR, WC, BMI and HCR was identified. The greater the area under the ROC curve, the greater the discriminatory power of indexes of obesity for HCR was. A confidence interval (CI) of 95% was used. CI at 95% calculation determines if the predictive ability of indexes of obesity is not due to chance, and its lower limit must not be lower than 0.50<sup>11</sup>. </p>
      <p>In the sequence, the sensibility and specificity were calculated, besides the cut-off points for C index, WHCR, WC, BMI and HCR. Values identified through ROC curve consist of cut-off points that must promote a more suitable balance between sensitivity and specificity for C index, WHCR, WC and BMI, as discriminators of HCR. </p>
      <p>Data were analyzed through the statistic program "STATA", version 7.0. </p>
      <p>
        <bold>Results</bold>
      </p>
      <p>The characteristics of the sample are demonstrated on <xref ref-type="table" rid="tab01-body2">table I</xref>. It is observed that despite not having statistically significant differences between men and women in the variables educational level, race and coronary risk, most of participants in the sample show low coronary risk and are not whites. Statistically significant differences are observed between men and women in the variables of body weight, height, waist and hip circumference, SBP, DBP, total cholesterol, HDL-C, C index, BMI, WHCR and WC. </p>
      <table-wrap id="tab01-body2">
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt01-en.jpg"/></table-wrap>
      <p> The areas under the ROC curve between C index, BMI, WHCR, WC and HCR, in people of male sex, can be observed in <xref ref-type="fig" rid="fig01-body2">figure I</xref>, noting that C index shows the highest percentage under the curve among the indexes of obesity to discriminate HCR in men, with statistically significant differences among them selves (<xref ref-type="fig" rid="fig01-body2">fig. 1</xref> and <xref ref-type="table" rid="tab02-body2">tab. II</xref>). </p>
      <fig id="fig01-body2">
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gf01-en.jpg"/></fig>
      <table-wrap id="tab02-body2">
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt02-en.jpg"/></table-wrap>
      <p>In <xref ref-type="fig" rid="fig02-body2">figure 2</xref>, it is possible to observe that, for women, the areas under the ROC curve between C index, BMI, WHCR, WC and HCR are similar for C index and WHCR, with statistically significant differences among the many indexes of obesity to discriminate HCR (<xref ref-type="fig" rid="fig02-body2">fig. 2</xref> and <xref ref-type="table" rid="tab03-body2">tab. III</xref>). </p>
      <fig id="fig02-body2">
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gf02-en.jpg"/></fig>
      <table-wrap id="tab03-body2">
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt03-en.jpg"/></table-wrap>
      <p>The values of sensitivity and specificity with more suitable balance among themselves, in addition to their respective cut-off points, are shown for all indexes of obesity as discriminators of HCR. It is noted that, regarding the male sex, C index is shown as the index of obesity with the best values of sensitivity and specificity to discriminate HCR. Among women, WHCR shows values discreetly greater than C index's (<xref ref-type="table" rid="tab04-body2">tab. IV</xref>). </p>
      <table-wrap id="tab04-body2">
        <graphic xlink:href="1678-4170-abc-85-01-0026-00006-gt04-en.jpg"/></table-wrap>
      <p>
        <bold>Discussion</bold>
      </p>
      <p>Concerning the cut-off points for WC identified in the present study, the results show divergences with literature. There is a proposal for vales of WC over 102cm (men) and 88cm (women) was better cut-off points to identify obesity and HCR<sup>12</sup>. In the analyzed sample, the use of such values would imply in a great number of people incorrectly classified, which would generate many false-negative, as the cut-off points identified in the participants in the study were 88cm for men and 83cm for women. </p>
      <p>When the comparison among the several indexes of obesity with HCR was made, it was observed that C index and WHCR are its best discriminators, despite the values of sensitivity and specificity found are not the ideal ones. There are some comparative studies among the many indexes of obesity as discriminators of HCR. However, studies that include the C index in this type of analysis were not found in the referred literature. </p>
      <p>In study carried out with 2,728 Japanese people of both sexes, the best cut-off points observed by the authors for BMI, WC and WHCR as discriminators of high cardiovascular risk among men were 23.5 kg/m&#178;, 84 cm and 0.90, whereas among women the cut-off points were 22.5 kg/m<sup>2</sup>, 72cm and 0.80, respectively. The authors concluded that the central adiposity indexes discriminate the HCR better than the total obesity indexes<sup>13</sup>. </p>
      <p>Our results showed some similarities with the previously mentioned study, especially in the cut-off points for BMI and WHCR among men, and WHCR among women. The cut-off point observed for WC as a discriminator of HCR among women in our study was much greater than the one shown among Japanese people. </p>
      <p>In a similar study<sup>14</sup>, performed with approximately 80,000 Chinese individuals, the best cut-off points for BMI and WC as discriminators of HCR were identified. For the BMI, the value of 24 kg/m&#178; was identified for both sexes, whereas for WC the value of 85 cm was identified for men and 80 cm for women. Those values were very close to those found in Salvador - Brazil. </p>
      <p>The discriminatory power of BMI and WC, as screening tools for cardiovascular risk was analyzed in 5,149 French women<sup>15</sup>, which found similar areas under ROC curve for both indexes of obesity, when used to discriminate women with at least one cardiovascular risk factor: 0.71 (0.69-0.73) and 0.76 (0.74-0.78) for BMI and WC, respectively. Despite the areas under the ROC curve show approximated values, the authors report that WC has a greater discriminatory power to classify HCR than BMI has. Our results demonstrate lower values in the areas under the ROC curve, probably due to the grouping of many cardiovascular risk factors in a single coronary risk index. The same authors found different areas under the ROC curve for each one of the cardiovascular risk factors when solely analyzed. </p>
      <p>In a study carried out in Taiwan<sup>16</sup> to determine the best cut-off points of many anthropometric variables as discriminators of cardiovascular risk, 26,359 men and 29,204 women were analyzed. The following cut-off points were found to determine at least one cardiovascular risk factor: BMI (23.6 and 22.1 kg/m), WC (80.5 and 71.5 cm) and WHCR (0.85 and 0.76) for men and women, respectively. </p>
      <p>The values of our cut-off points were higher, which demonstrates that in Taiwan's population the behavior of anthropometric variables as discriminators of HCR shows different characteristics from our population's. </p>
      <p>Other studies were carried out in the perspective of identifying cut-off points and compare the many indexes of obesity as discriminators of cardiovascular risk factors individually analyzed. In India<sup>17</sup>, the values of 85 cm for men and 80 cm for women, for WC, and 0.88 (men) and 0.81 (women) for WHCR were identified as cut-off points to discriminate diabetes. </p>
      <p>In another work, with the aim of comparing C index and BMI was hyperglycemia discriminators, in 1,325 adults in the city of Salvador-BA, C index was found with a greater discriminatory power than BMI, with areas under the ROC curve of 0.71 (0.66-0.77) and 0.62 (0.57-0.67), respectively<sup>18</sup>. </p>
      <p>In another study<sup>19</sup>, using the same population, C index and BMI were compared as systolic hypertension, finding areas under ROC curve without statistically significant differences, showing that the discriminatory power for HCR between C index and BMI is similar. </p>
      <p>One of the limitations of the present study was the HCR classification using the proposed algorithm based on the cohort of Framinghan<sup>9</sup>. We used that algorithm by applying its values in the individuals participating in our sample and define that HCR would be identified through points refereeing to the percentile 80 in the percentile distribution of the study population itself. There are controversies concerning the ideal points for identification of HCR in a short term. There are also suggestions<sup>20</sup> that the HCR could be identified through the points equivalent to 20% from absolute risk for coronariopathy in the next 10 years, which would be equivalent to 9 points for men and 15 points for women, in the proposed algorithm based on the cohort of Framinghan<sup>9</sup>. In our study, HCR was identified through the cut-off point of 8 points for men and 10 points for women. </p>
      <p>Recent studies have demonstrated that the long-term prediction of HCR, done through the proposed algorithm based on the cohort of Framinghan<sup>9</sup>, can be generalized for other populations, as well as used in cross-section studies. The generalization for other populations must be done after re-gauging for different prevalences of risk factors involved in the analysis<sup>21, 22</sup>. </p>
      <p>Concerning the variables used in the proposed algorithm, the cut-off point adopted for diabetes was the value of 140mg/dl, whereas in this study the value was 126mg/dl, considering the recent suggestions from many scientific associations<sup>23, 19, 24</sup>. </p>
      <p>Many other variables can be included in predictive models of coronary risk. Early familial history of CAD, despite being strongly associated to coronary events, is not available in the subjects analyzed in the proposed algorithm based on the cohort of Framinghan<sup>8</sup>. The fibrinogen has been associated to cardiovascular risk<sup>23</sup>, but it is not also available in the previously mentioned study<sup>8</sup>. Besides, due to its high biological variability and absence of corroboration that specific interventions decrease the cardiovascular risk, its routine determination is not recommended <sup>23</sup>. Left ventricular hypertrophy, observed through ECG, has been used in algorithms for prediction of CAD. However, considering its high association with hypertension, it was not used in the previously mentioned algorithm<sup>8</sup>.</p>
      <p>On the other hand, when the comparison of different analysis methods of HCR was made, the incorporation of total cholesterol/HDL-C ratio in the algorithms for classification of HCR<sup>25</sup> is suggested. In our study, the total of points of total cholesterol separately from HDL-C was used. </p>
      <p>According to the results found, there were evidences that the central obesity indexes, especially C index and WHCR, are better to discriminate HCR than indexes of general obesity (BMI). WC was a reasonable performance, despite many studies have demonstrated that it is a powerful discriminator of HCR (data were not confirmed in our observations). Those results suggest that abdominal obesity indexes are better to discriminate HCR than the indexes of general obesity. New studies with infants, adolescents and adults from different age groups are suggested to identify and compare the discriminatory power of anthropometric indexes of obesity and HCR in such specific subgroups.</p>
      <p>
        <bold>References</bold>
      </p>
      <p>
        <fn id="end-body2" fn-type="corresp">
          <label>
            <bold>Correspondence to</bold>
          </label>
          <p> Francisco Jos&#233; Gondim Pitanga </p>
          <p> Av. Luiz Tarqu&#237;nio Pontes, 600 </p>
          <p> 42700-000 - Lauro de Freitas, BA - Brazil</p>
          <p> E-mail: </p>
          <p>
            <email xlink:href="pitanga@lognet.com.br">pitanga@lognet.com.br</email>
          </p>
        </fn>
      </p>
      <p>Received for publishing on 05/11/2004 </p>
      <p> Accepted on 01/21/2005</p>
    </body>
  </sub-article>
</article>
