TESE Final
TESE Final
TESE Final
Belo Horizonte - MG
2020
Marcelo Martins Pinto Filho
Belo Horizonte - MG
2020
Pinto Filho, Marcelo Martins.
P659a Alterações maiores do eletrocardiograma na população brasileira: valor
prognóstico e determinantes de incidência - Estudo Longitudinal de Saúde
do Adulto (ELSA-Brasil) [manuscrito]. / Marcelo Martins Pinto Filho. - - Belo
Horizonte: 2020.
97 f.: il.
Orientador (a): Antonio Luiz Pinho Ribeiro.
Coorientador (a): Luisa Campos Caldeira Brant.
Área de concentração: Infectologia e Medicina Tropical.
Tese (doutorado): Universidade Federal de Minas Gerais, Faculdade de
Medicina.
NLM: WG 140
Fabiene Letízia Alves Furtado | Bibliotecária CRB-6/2745
UNIVERSIDADE FEDERAL DE MINAS GERAIS
FACULDADE DE MEDICINA
Diretor: Prof. Humberto José Alves
Vice Diretora: Profa. Alamanda Kfoury Pereira
( ) Aprovada
( ) Reprovada
Finalizados os trabalhos, lavrei a presente ata que, lida e aprovada, vai assinada pelos membros
da Comissão.
Ao Prof. Antonio Luiz Pinho Ribeiro (Tom), professor, mentor, exemplo de lucidez,
competência e brilhantismo científico. Das conversas musicais às científicas, pude evoluir
como pesquisador, médico e ser humano.
À Profª. Luisa Campos Caldeira Brant que despertou em mim o interesse pela
epidemiologia em cardiologia e que combina com maestria pesquisa e assistência.
À Profª. Graziela Chequer (Grazi), colega de enfermaria e amiga de prosa. Entra ano e sai
ano, entra artigo e sai artigo, continua referência na medicina e na arte.
Ao Marco Paulo Tomaz Barbosa, quem me ensinou o caminho para a prática baseada em
evidências. Anos passados da residência em cardiologia, continua meu amigo e preceptor.
ARTIGO 1
Table 1: Characteristics of the Brazilian Longitudinal Study of Adult Health participants
with valid electrocardiogram at baseline (general population and stratified by presence of
major cardiovascular abnormalities according to Minnesota coding criteria) (n= 13.428). P
.. 61
Table 2: Prevalence of major cardiovascular abnormalities and their respective crude hazard
ratios and 95% confidence interval, for the outcome of total mortality (n= 13.428) .. .. 62
Table 3: Total mortality rates and hazard ratio according to the presence of major
electrocardiographic abnormalities . 62
Table 4: Systematic Coronary Risk Evaluation score reclassification for the outcome of
ardiographic
. 63
Supplemental table 1: Total mortality rates and hazard ratios according to sex and presence
of major electrocardiographic abnormalities 67
ARTIGO 2
Table 1: Characteristics of the Brazilian Longitudinal Study of Adult Health participants
with valid electrocardiogram at baseline stratified by sex and development of major
cardiovascular abnormalities according to Minnesota coding criteria (n= 11. ..........81
Table 2: Incidence of specific and total new major electrocardiographic abnormalities
measured at 82
LISTA DE FIGURAS
Figura 1: Mortalidade Global por doenças cardiovasculares entre 2007 2017 ................ 17
Figura 2: Mortes a cada 100 mil habitantes em 2017, por doença arterial coronariana (em
ambos os sexos e todas as idades) ........................................................................................ 18
Figura 3: Principais causas de morte no Brasil em 1990 e 2017, para todas as idades em
ambos os sexos .................................................................................................................... 19
Figura 4: Traçado eletrocardiográfico normal .................................................................... 22
ARTIGO 1
Figure 3: Fine and Gray model for competitive risk from non-cardiovascular and
cardiovascular death between those with and without major electrocardiographic
a ... 66
Supplemental figure 2: Receiver operating characteristic curves for the SCORE risk
calculator for cardiovascular mortality and for the modified risk score including major
... ... 69
ARTIGO 2
Figure 1: .................................... 80
LISTA DE ABREVIATURAS E SIGLAS
CM - Código de Minnesota
CV - Cardiovascular
FA - Fibrilação atrial
HAS - Hipertensão arterial
DM - Diabetes mellitus
FC - Frequência cardíaca
OR - Odds ratio
IC - Intervalo de confiaça
HR - Hazard ratio
1. CONSIDERAÇÕES INICIAIS........................................................................................ 14
2. INTRODUÇÃO .............................................................................................................. 14
4 OBJETIVOS .................................................................................................................... 35
4.1 Objetivos Gerais ............................................................................................................ 35
4.2 Objetivos Específicos .................................................................................................... 35
5 MATERIAIS E MÉTODOS............................................................................................. 36
5.1 Participantes .................................................................................................................. 36
5.2 Variáveis ....................................................................................................................... 37
5.3 Obtenção do ECG........................................................................................................... 38
5.4 Determinação de Eventos e Desfechos .......................................................................... 41
5.5 Análise estatística .......................................................................................................... 42
5.6 Aspectos Éticos.............................................................................................................. 43
5.7 Critérios da Pesquisa bibliográfica ................................................................................ 43
6 RESULTADOS.................................................................................................................45
6.1 Artigo 1 submetido na .................................................................................................... 45
6.2 Artigo 2 ainda não submetido ........................................................................................ 70
7 CONSIDERAÇÕES GERAIS.......................................................................................... 83
8 CONCLUSÕES................................................................................................................ 84
1. CONSIDERAÇÕES INICIAIS
2. INTRODUÇÃO
pública (1, 5). Vinculada a alta prevalência dessas doenças, está a grande possibilidade de
prevenção. Estima-se que até 80% das DCV no mundo sejam evitáveis (6, 7). Isso significa,
na prática, imenso potencial de melhoria na qualidade de vida e prevenção de mortes
globalmente.
No sentido da prevenção, é fundamental, e cabe a todos, orientações para que os
indivíduos busquem um estilo de vida saudável, pratiquem exercício físico e mantenham
peso e dieta adequados (6, 8). Entretanto, pessoas definidas como de alto risco cardiovascular
(CV) devem ser candidatas a controle mais agressivo dos fatores de risco. Isso inclui
orientações mais enfáticas sobre hábitos de vida saudáveis e, até mesmo, o uso de
medicamentos com impacto na redução do risco cardiovascular (a exemplo das estatinas).
Nesse sentido, é desejável que possamos estimar, da maneira mais acurada possível, o risco
CV de um indivíduo, para assim identificar corretamente aqueles de alto e de baixo risco
CV. Essa estimativa é pilar fundamental da abordagem ao paciente em que planejamos
estratégias de prevenção primária das DCV (8-10).
Para o cálculo de risco, utilizamos os escores de risco CV. Ferramentas bem
validados em coortes internacionais, que levam em conta para sua construção, fatores de
risco tradicionais para DCV. Como exemplos, têm-se o Framinghan Risk Score (11), o
Pooled Cohort Equation (PCE) (10) e o europeu Systematic Coronary Risk Evaluation score
(SCORE) (12), que classificam os pacientes em risco baixo, intermediário e alto.
Sabe-se que pacientes considerados de baixo risco, devem ser orientados quanto a
hábitos de vida saudáveis. Os indivíduos de alto risco, além dessas medidas, devem ter o
tratamento com hipolipemiantes oferecidos (8, 9). Aos sujeitos ditos de risco moderado,
entende-se o potencial benefício da intervenção farmacológica, porém com menor
magnitude de benefício do que quando comparados aos indivíduos de alto risco (13). Dessa
maneira, uma estimativa de risco acurada é imperativa para uma boa estratégia de prevenção
de DCV.
Esses escores de risco atualmente utilizados possuem acurácia moderada e, em geral,
superestimam o risco individual, com uma estatística C girando em torno de 0,70 (14, 15).
Alguns fatores de risco CV não contemplados nesses escores, como a história familiar de
DCV, ou um escore de cálcio elevado, são potenciais modificadores de risco, e podem ajudar
no manejo individualizado do paciente (8, 9, 16). Nesse sentido, ferramentas adicionais aos
escores de risco tradicionais, podem ajudar a melhor definir o prognóstico e risco de um
indivíduo. A diretriz da Sociedade Europeia de Cardiologia de 2016 já ressalta a importância
16
3. REVISÃO DA LITERATURA
Figura 1: Taxa de Mortalidade Bruta Global por Doenças Cardiovasculares entre 2007
2017.
Fonte: Institute for Health, Metrics and Evaluation. Disponível em Evaluation IHME. Findings from the Global
Burden of Disease Study 2017. The Lancet: IHME; 2018.
18
Figura 2: Taxa de mortalidade ajustada por idade por 100 mil habitantes em 2017, por
doença arterial coronariana.
Figura 3: Principais causas de morte no Brasil em 1990 e 2017, para todas as idades em
ambos os sexos.
O impacto social das DCV é tremendo. No Brasil, elas são a principal causa de
redução nos anos saudáveis de vidada população ( sability-adjusted life - DALY),
sendo responsáveis por aproximadamente 14% do total de DALY. Esse valor supera, pela
sua alta prevalência, até mesmo a perda de DALY por violência, mortes de trânsito e doenças
do período neonatal, que tipicamente acometem indivíduos mais jovens (37). Soma-se a isso
o alto custo para o tratamento dessas doenças. Estima-se que nos EUA, em 2035 o custo com
o tratamento das DCV poderá atingir a impressionante cifra de 1,1 trilhão de dólares (USD)
(38). No Brasil, a estimativa atual de custo gira em torno de 26 bilhões de USD (39).
Por essas questões, e levando-se em conta o enorme potencial de prevenção das DCV
por meio do controle de fatores de risco (6, 7), torna-se imperativo que esforços sejam
dirigidos para estratégias de redução da carga das DCV no mundo moderno. Porém, a melhor
20
3.3 O eletrocardiograma
O ECG basal de 12 derivações é um dos exames mais frequentemente realizados na
prática médica. É o teste fundamental para a análise de ritmo, do sistema de condução
cardíaco e para avaliação de isquemia miocárdica. Mesmo com o extenso desenvolvimento
tecnológico na propedêutica CV, o ECG ainda tem papel central na abordagem de sintomas
importantes como dor torácica e síncope. É também extremamente útil na detecção de
distúrbios hidroeletrolíticos, na monitorização do uso de alguns medicamentos e adjuvante
no diagnóstico de doenças valvares e do pericárdio (45). Fundamentalmente, se traduz em
um gráfico de voltagem em seu eixo vertical (registrada pela diferença de potenciais elétricos
entre duas regiões do coração), pelo tempo em seu eixo horizontal (46).
Fonte: https://meds.queensu.ca/central/assets/modules/ts-ecg/normal_ecg.html
Estabelecendo-se condições para a realização do ECG nesse tipo de estudo, seu uso
foi se tornando mais frequente. O valor das alterações eletrocardiográficas na predição de
eventos cardíacos maiores [infarto do miocárdio (IAM), acidente vascular encefálico (AVC)
e morte de causa cardiovascular] pode ser mais bem avaliado nas décadas seguintes. Em
1978 foi publicada uma meta-análise envolvendo oito estudos populacionais iniciais, com
um total de 8390 indivíduos do sexo masculino, em seguimento médio de oito anos (59). A
importância das anormalidades ao ECG como marcadores de doença cardíaca subclínica e
preditores de evento cardíaco maior pode ser confirmada. Mais recentemente outras coortes,
já realizadas no contexto atual de melhoria na estratégia de prevenção primária de DCV
incluindo a prevenção medicamentosa com hipolipemiantes (tendo as estatinas como classe
principal), também confirmaram o poder do ECG em predizer eventos CV (18, 24, 60-66).
permanece essencialmente inalterado desde sua criação. Apenas recentemente, houve uma
tentativa de se produzir uma abordagem alternativa, feita por Rautaharju (um dos autores do
Código de Minnesota original) e colaboradores (56).
É digno de nota, que tais alterações se tornam mais prevalentes a medida que os
participantes acumulam fatores de risco para DCV, tanto em homens quanto em mulheres,
como ilustra a figura 5. Na figura 6, podemos perceber, que a prevalência das AEM também
aumenta com a idade.
27
Fonte: Major Electrocardiographic Abnormalities According to the Minnesota Coding System Among Brazilian
Adults (from the ELSA-Brasil Cohort Study), Am J Cardiol, Pinto-Filho et al.
28
Fonte: Major Electrocardiographic Abnormalities According to the Minnesota Coding System Among Brazilian
Adults (from the ELSA-Brasil Cohort Study), Am J Cardiol, Pinto-Filho et al.
Ondas Q maiores 2.738 11,2 11,6 20,4 6,9 14,4 12,1 17,6 7,8 13,8
(MC 1.1, 1.2)
Alterações de ST 1.952 6,8 4,9 9,6 4,7 10,6 9,3 16,5 7,0 15,0
maiores (MC 4.1,
4.2)
Alterações de T 5.234 19,1 11,3 24,3 16,8 29,7 21,4 33,5 21,4 34,4
maiores (MC 5.1,
5.2)
Alterações de 998 3,9 2,6 14,9 1,0 5,8 3,6 12,2 1,6 4,4
condução AV
(MC 6.1, 6.2, 6.4,
6.8)
Bloqueio de ramo 293 1,2 0,3 2,7 0,6 3,1 0,4 1,1 0,3 2,4
esquerdo (MC 7.1)
Bloqeuio de ramo 726 2,7 2,8 9,4 0,6 3,8 1,8 7,4 0,7 4,4
direito (MC 7.2)
Bloqueio 140 0,6 0,6 2,3 0,1 0,9 0,5 1,3 0,1 0,4
intraventricular
inespecífico (MC
7.4)
Fibrilação atrial (MC 277 1,2 0,4 5,2 0,2 2,6 1,0 1,8 0,02 1,0
8.3)
Ritmo 116 0,4 0,2 0,9 0,2 0,7 0,2 0,7 0,4 0,7
supraventricular
(MC 8.4)
Total 6.790 28,2 23,2 37,1 25,4 35,7 28,8 37,4 28,4 34,9
MC: Código de Minnesota; AV: atrioventricular
Modificado de: Prineas et al. United States national prevalence of electrocardiographic abnormalities in black
and white middle-age (45- to 64-Year) and older (>/=65-Year) adults (from the Reasons for Geographic and
Racial Differences in Stroke Study) Am J Cardiol.
(principalmente maiores) pelo CM. Para uma melhora compreensão da relação temporal
entre elas, estudos de incidência de AEM também devem ser explorados.
com HbA1c < 8% (15,8 versus 10,5%). Outros fatores de risco também associados a
ocorrência de AEM novas, mesmo após ajuste por modelos multivariados foram tabagismo:
Hazard Ratio (HR) 1,75 (IC 95% 1,22-2,53) e pressão arterial sistólica: HR 1,03 (IC 95%
1,01 1,05).
eventos não fatais. A derivação desses dados foi validada internamente por meio de
bootstraping (62). Por outro lado, Groot e colaboradores, em uma coorte holandesa (44),
conseguiram uma melhora no NRI apenas modesta e não significativa, ao incorporar sobre
o SCORE europeu, a presença de alterações eletrocardiográficas predeterminadas (IAM e
FA). O NRI obtido foi de apenas 1,0% para IAM ao ECG (IC 95% 3,2-6,9) e 0,5% para FA
(IC 95% -3,5-3,3) sem melhoria na estatística C obtida.
32
2012 Auer e 2192 homens e AEM e alterações Mortalidade por Aumento de risco para
colabora-dores mulheres entre 70 menores, pelo CM qualquer causa doença coronariana após
a 79 anos de idade, ou evento modelo multivariado. HR
(74).
negros e brancos, coronariano 1,35(IC 95% 1,02-1,81)
americanos, follow incidente para alterações menores e
up de 6,4 anos HR 1,51(IC 95% 1,20-
1,90) para AEM
2013 Badheka 8.527 homens e Duração do complexo Mortalidade CV BRE, BRD foram
e mulheres QRS, definida em associados
americanos, com quartis (61-89/90-97/98- independentemente ao
colaboradores idade média de 105/106-209) desfecho de interesse.
(21) 60,5 anos, com Independente da
follow up de 12,5 morfologia, o quartil com
anos maior duração de QRS
apresentou maior
mortalidade do que o
menor quartil (HR 1,3
IC95% 1,01-1,7)
2014 Shafiee e 247 participantes Alterações DCV e Em modelo multivariado,
colaboradores do Iran, idade eletrocardiográficas mortalidade por AEM estiveram associadas
média de 76,7 maiores pelo CM qualquer causa ao desfecho de interesse
(18). anos, folow up 3.2 HR 3,12 (IC 95% 1,02-
anos 9,57) para morte CV, HR
2,45 (IC 95% 1,23-4,85)
para mortalidade geral
2016 Rezaeian 6.765 participantes Alterações Mortalidade por As AEM e menores
e colaboradores assintomáticos, eletrocardiográficas DCV, IAM, AVE estiveram associadas aos
homens e maiores ou menores ou ICC novas. desfechos de interesse, de
(17) mulheres, pelo CM maneira independente,
americanos, de após ajustes por modelo
raças diversas, multivariado.
idade média de 60
anos.
EUA: Estados Unidos da América; DM2: Diabetes Mellitus tipo 2; CM: código de Minnesota; AEM: alterações
eletrocardiográficas maiores; HR: hazard ratio; IC: intervalo de confiança; HVE: hipertrofia ventricular esquerda; DCV:
doença cardiovascular; IAM: infarto agudo do miocárdio; ECG: eletrocardiograma; AVE: acidente vascular encefálico;
CV: cardiovascular; BRE: bloqueio de ramo esquerdo; BRD: bloqueio de ramo direito; IC: insuficiência cardíaca
alterações clinicamente conhecidas. São elas: IAM definido pelo ECG, HVE, FA, BRE.
BRD, alterações do segmento ST-T, BAV de segundo e terceiro grau. Além disso, este
estudo recrutou um número menor de participantes (2.370), com uma média de idade menor
(48 anos). Nesse cenário, o ECG não teve poder para reclassificar os participantes. Em
publicação no JAMA Cardiology em 2016 (43) Shah e colaboradores realizaram uma
abordagem diferente: utilizaram mensurações eletrocardiográficas (eixo das ondas P, R, e T;
FC; duração do complexo QRS; e duração dos intervalor PR e QT) e concluíram, nessa
34
coorte de 3640 participantes recrutados entre 1971 e 1975, com idade média de 55 anos, que,
ao incluir as variáveis: FC, eixo de onda T e intervalo QT, ao escore de Framingham, obtêm-
se melhoria estatisticamente significativa na estatística C (de 0,76 para 0,80). Tais achados
passaram por validação externa em outra coorte, sendo confirmados.
as DCV nas diferentes raças, e em ambos sexos, é fundamental obter informações que, de
maneira mais confiável, reflita o perfil da população brasileira. O viés da raça impossibilita
que os achados em estudos anteriores sejam aplicados, sem ressalvas, em países com origem
étnica/racial substancialmente diferente. Até o presente momento, não encontramos estudos
referentes ao valor prognóstico das alterações eletrocardiográficas no Brasil ou na América
Latina.
A repetição dos exames realizados na linha de base (incluindo o ECG) possibilita ainda
a análise de determinantes do surgimento de novas AEM.
4. OBJETIVOS
5.1 Participantes
5.2 Variáveis
alto risco pelo SCORE são mais quando comparados aos riscos derivados das coortes
38
A coleta de dados se deu nos Centros de Investigação ELSA. Nas onda 1 e 2, foram
obtidos traçados de ECG de 12 derivações utilizando o aparelho digital (Atria 6100, Burdick,
Cardiac Science Corporation, Bothel, WA, USA) com leitura automatizada da frequência
cardíaca, duração, amplitude e eixos das ondas P, QRS e T, além dos intervalos QT, QTc e
da dispersão de QT. Os eletrocardiogramas foram transmitidos do eletrocardiógrafo Atria
6100 para um servidor do Centro de Leitura em Eletrocardiografia, localizado no CL-
ELSA/MG, onde foram armazenados para posterior análise em sistema de gerenciamento de
eletrocardiogramas Pyramis (versão 6.2.b, Cardiac Science Corporation, Bothel, WA, USA,
Figura 7) (33, 52), com possibilidade de edição do traçado e mensuração dos intervalos
batimento a batimento. As definições de alterações maiores e menores foram feitas através
do CM, já validado internacionalmente com revisão manual dos ECG codificados com
códigos compatíveis com alterações eletrocardiográficas maiores (50, 52). A leitura e
classificação dos eletrocardiogramas pelo CM, como já descrito anteriormente, foi
centralizada no ELSA MG.
Na ocasião da terceira onda de entrevista e exames do ELSA-Brasil, iniciada em 2017,
já com nove anos de condução do estudo, o processo de obtenção e transmissão do ECG foi
atualizado para fins de melhoria na eficácia da transmissão e armazenamento dos exames. O
eletrocardiógrafo TEB (tecnologia eletrônica brasileira) ECGPC® passou a ser utilizado, por
39
prover comunicação direta com a rede de telessaúde do Hospital das Clínicas da UFMG,
aonde está montado o centro de leitura de ECG. A telessaúde do Hospital das Clínicas utiliza
uma plataforma própria de telediagnóstico, do Sistema Integrado de Gestão de Serviços de
Telessaúde (SIGTEL) (figura 8). Com a transmissão direta e automática para essa
plataforma, o risco de perda de exames pode ser reduzido para virtualmente zero. Soma-se
a isso a maior facilidade e praticidade de visualização do traçado e realização de medidas
pelo SIGTEL, justificando, portanto, a mudança. Apesar da troca do eletrocardiógrafo
utilizado, a técnica de obtenção dos traçados foi a mesma, assim como se manteve a mesma
estrutura de coordenação e leitura dos ECG. Os traçados obtidos nessa terceira onda ainda
estão passando por processo de análise final, não estando disponíveis no momento.
Os ECG obtidos na onda 1 já haviam sido codificados e analisados, tendo sido alvo de
publicações prévias (34, 35). Os ECG das ondas seguintes, foram revisados manualmente
para a codificação pelo CM, caso tivessem apresentados AEM discordantes dos ECG da
onda 1. Os participantes com pelo menos uma alteração maior pelo código de Minnesota
(independentemente da presença de alterações menores), foram classificados como:
presença de alteração maior.
A obtenção de traçados eletrocardiográficos em estudos epidemiológicos
multicêntricos requer padronização. É imperativo que se determine orientações bem
precisas, a fim de se evitar variações entre examinadores e perda de traçados ou erros em
sua obtenção. No caso do estudo ELSA, profissionais foram capacitados especificamente
para a obtenção dos ECG. Após passarem por avaliações de qualidade foram licenciados
para a realização dos exames (33).
A obtenção dos traçados se deu nos Centros de Investigação, no período da manhã. Os
participantes deveriam estar em um ambiente confortável, abstinentes de álcool e cigarro nas
24 horas que antecederam o exame que seguia uma rotina específica sugerida pelo manual
do aparelho.
Os procedimentos para a realização dos exames foram:
-Alocação do participante em um ambiente confortável
41
doenças (CID 10), utilizando-se registros do grupo I, especificamente I20-I25, I50, I60-I69
que tratam de doenças cardiovasculares e cerebrovasculares.
As definições e classificações dos desfechos considerados de natureza cardiovascular,
se deram de acordo com critérios já validados internacionalmente (79, 80). A equipe de
avaliação de desfechos cardiovasculares, composta por cardiologistas ficou encarregada da
adjudicação dos eventos suspeitos de: Síndrome coronariana aguda (SCA), ICC,
revascularizações arteriais, doença arterial periférica, trombose venosa profunda,
tromboembolismo pulmonar e FA. Os eventos relacionados a doença cerebrovascular foram
adjudicados por uma equipe de neurologistas e o processo ainda está em andamento. Para o
presente estudo, utilizamos como desfecho: mortalidade global, mortalidade CV e os
seguintes desfechos CV não fatais: SCA, ICC, revascularização coronariana.
SCORE. Além disso, em análise preliminar, o aumento de risco encontrado pela presença
das AEM pelo ECG foi semelhante nos dois sexos.
Para a análise do desfecho primário, foram construídas curvas de Kaplan-Meir,
comparadas pelo teste de log rank. A força de associação foi medida pelo modelo de Cox,
com a determinação da razão de risco (HR), após ajustes por idade, raça, escolaridade e
fatores de risco para DCV. O pressuposto de proporcionalidade foi avaliado pela análise
visual dos gráficos e telo teste de Schoenfeld. Para avaliação de mortalidade específica
usamos o modelo de Fine e Gray de risco competitivo. Para a análise de determinantes da
incidência de alterações eletrocardiográficas maiores, os grupos foram comparados por meio
43
de regressão logística. Definimos que este seria o teste mais apropriado, já que temos apenas
2 medidas separadas por uma média de 4 anos e não um seguimento linear do ECG, não nos
permitindo definir o tempo passado até o surgimento de cada alteração
O cálculo do escore de risco pelo SCORE foi realizado para todos os participantes e
ajustado para o desfecho de interesse (mortalidade geral) pelo modelo de Cox. Para avaliar
se a incorporação das variáveis eletrocardiográficas de interesse (AEM pelo CM) ao escore
de risco resultaria em um modelo preditivo de maior acurácia, foi realizado o cálculo do NRI
e da estatística C para o novo modelo proposto. A princípio todas as AEM seriam elegíveis
a serem testadas no modelo.
assim consideradas diferenças estatisticamente significativas aquelas cujo p-valor foi menor
que 0,05, exceto se explicitado de outra maneira.
Para a análise dos dados, foi utilizado o pacote estatístico IBM SPSS Statistics (IBM
Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM
Corp) e o R (R Core Team 2013. R: A language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/).
Marcelo M. Pinto-Filho, MD, MSca,b, Luisa C. C. Brant, MD, PhDa,b, Rodrigo C P dos Reis,
PhDc, Luana Giatti, MD, PhDa, Bruce B Duncan, MD, PhDc,d, Paulo A. Lotufo, MD, PhDe,f,
Maria de Jesus M. da Fonseca, PhDg, José G. Mill, MD, PhDh, Maria da Conceição C. de
Almeida, PhDi, Peter W. Macfarlane, DSc, FRCPj, Sandhi M. Barreto, MD, PhDa, Antonio
Hospital Borges da Costa. Alfredo Balena Av, 110 Funcionários ZIP 30130-100 Belo
ABSTRACT
METHODS: Participants from the Brazilian Longitudinal Study of Adult Health (ELSA-
Brasil), free of known CVD at baseline were included. A 12-lead ECG was obtained at
baseline (2008-2010). Participants were followed up to 2018 by annual interviews. Deaths
and hospitalizations were independently reviewed. Cox as well as Fine and Gray multivariate
regression models were applied to evaluate if the presence of any major electrocardiographic
abnormality (MEA) at baseline, defined according to the Minnesota Code system, would
predict total and cardiovascular deaths. We also evaluated the impact of adding MEA to the
Systematic Coronary Risk Evaluation (SCORE) risk prediction model for cardiovascular
mortality, using C-statistic and the net reclassification index (NRI).
RESULTS: The 13,428 participants (median age 51 years, 45% men) were followed for 8±1
years. Baseline prevalence of MEA was 6.4%. All-cause and cardiovascular mortality
occurred in 2.8% and 1.2% of the population, respectively. Prevalent MEA was an
independent predictor of overall mortality after adjustments for age, race, education and
traditional cardiovascular risk factors (HR:2.3; 95% CI 1.7-2.9). In the competitive risk
model for cardiovascular death, the increased risk attributable to MEA was even higher
(HR=4.6, 95% CI 3.0-7.0). Adding MEA to the SCORE resulted in 3% up miss
reclassification in the non-event subgroup and 15% correct up reclassification in those with
a fatal cardiovascular event. Absence of MEA did not result in down reclassification in any
subgroup.
CONCLUSION: Presence of MEA was a strong and independent predictor of overall and
cardiovascular mortality. ECG may have a role in risk prediction for death in primary care.
prediction, mortality
47
INTRODUCTION
Risk factors modification through diet, exercise, avoidance of tobacco exposure and
blood pressure control, should be advised for all that seek counselling for a healthier lifestyle
(3, 5). Nevertheless, a more aggressive approach, including pharmacological intervention, is
warranted for individuals with higher risk of a cardiovascular event. In addition to lifestyle
modifications, medical intervention with lipid lowering medications should also be offered
for those considered to be of high risk for CVD (3). Still, a large number of patients are
considered to be at intermediate risk. For these, treatment decisions are more challenging
since expected benefit from medical interventions is lower (7).
The electrocardiogram (ECG) is an inexpensive and widely available exam and the
presence of ECG abnormalities has been demonstrated to impact cardiovascular prognosis
(12, 13). Even today, a simple and useful strategy to use the ECG in cardiovascular risk
assessment is not yet available, and the US Preventive Services Taskforce classify the ECG
use for heart disease screening as uncertain for those considered at intermediate to high risk
(14). Using data from the Longitudinal Study of Adult Health (ELSA-Brasil), we aim to
evaluate if major electrocardiographic abnormalities (MEA), classified according to the
Minnesota Code system (MC), are independent predictors of all-cause and cardiovascular
mortality. We also investigated if MEA would improve risk assessment of the European
SCORE that also encompass only fatal cardiovascular events as outcomes of interest.
METHODS
The ELSA-Brasil cohort comprises 15,105 civil servants of both sexes, aged 35 to 74
years old, living in six Brazilian state capitals (Belo Horizonte, Rio de Janeiro, São Paulo,
Porto Alegre, Salvador and Vitoria). ELSA_Brasil aims to investigate the determinants of
CVD and diabetes. The study was approved by local ethics committees of the six
participating institutions. A detailed description of the study design and cohort profile was
subject of previous publication (15).
Study Variables
The variables of interest were those related to traditional cardiovascular risk factors,
potential risk modifiers and the cardiovascular risk estimated by a risk calculator.
Cardiovascular risk of death was calculated with the SCORE. The reason for choosing the
SCORE as the tool for risk estimation is because as in the original SCORE derivation model,
our paper will report only on fatal cardiovascular events. This risk score differs slightly for
different European countries. In general, eastern European countries are considered to have
a high cardiovascular baseline risk and western a low baseline risk. For its development in
our cohort we considered that it would be suitable to choose a low risk starting point (6).
Participants with a SCORE <1% are considered low risk, from 1 to <5% intermediate risk
and 5% high risk (8). Hypertension (no/yes) was defined as a systolic blood pressure (SBP)
(17) as well as self-declared use
of anti-hypertensive medication. Diabetes know
receiving treatment for di , post prandial glycaemia
mg/dl, .5% (18). In this paper, presence of dyslipidemia (no/yes)
was defined
-lowering medication we acknowledge that there
is variation in the definition of dyslipidemia among societies (19). Obesity (no/yes) was
smoki
primary education attained (in Brazil considered to be up to the fifth year of formal
education) were considered to be of low educational status. Another way we assessed
education was comparing those with college education to those without it. Skin color was
self-reported according to the Brazilian National Census (white, mixed/brown, black, Asian
and indigenous) (20).
Participants had their ECG recorded in the six different research centers according to
a standardized protocol (21), using the Burdick Atria 6100 machine, with a paper speed of
25mm/second and a calibration of 10mm/mV. Digital data were acquired with simultaneous
acquisition of 12 high-frequency leads (500 samples/second per channel), high-resolution
thermal print in 10 seconds. The recordings were sent to the reading center in Minas Gerais
Investigation Center for automated analysis using the Glasgow software (22). Coding was
automated according to the Minnesota Code System with manual review of the codes
50
associated to MEA by and ECG expert according to the ECG reading center routine. The
ECG findings at baseline have been described more comprehensively in previous
publications (23).
The MC provides classification of the abnormalities into minor and major categories
(24). Major electrocardiographic abnormality Minnesota codes were reviewed by a senior
cardiologist as standard practice in ECG reading centers. For the present study, only major
of major
abnormalities per participant.
The outcomes of interest for the study were total and cardiovascular mortality
(defined as death attributable to myocardial ischemia and infarction, heart
failure, cardiac arrest because of other or unknown cause, or stroke). Cases were identified
by annual telephone interview, hospital record review and report by the participants. Specific
cause mortality was defined by blinded adjudication of the underlying cause by a team of
cardiology specialists with at least two independent reviewers.
Statistical Analyses
For the survival analysis, individuals were followed up until death, loss to follow-up
or end of present analysis (31st Dec 2018), whichever came first. Kaplan-Meir curves were
built for the general population and according to pre-defined SCORE risk categories.
Survival curves were compared by the log rank test. To evaluate additional risk imposed by
the presence of MEA, a Cox regression model was built stratified by sex, and adjusted for
age, race, education and presence of traditional cardiovascular risk factors. The proportional
hazard assumption was tested by visual analyses of the curves
To evaluate the impact of MEA in cardiovascular mortality the Fine and Gray model for
competitive risk was utilized applying the same multivariate analysis used in Cox regression.
Cardiovascular risk was calculated with the SCORE (8). To evaluate if incorporating
MEA in the SCORE would increase the accuracy of the predictive model, we calculated the
net reclassification index (NRI) and the C statistic of the new proposed model. As a first
step, an initial regression was performed with adjustment for the calculated risk with the
SCORE. Thereafter a new risk was obtained for the participants multiplying the baseline
SCORE risk of individual patients by the hazard ratio found for the presence of MEA. The
new calculated risk was used to reclassify the individual in low-, intermediate- or high-risk,
according to the cut offs previously defined by SCORE. Subsequently, the proportion of
individuals reclassified according to cardiovascular mortality status was evaluated and
stratified according to correct and incorrect reclassifications Significance level was set at
0.05 unless stated otherwise. We used IBM SPSS Statistics 20.0 (IBM Corp. Released 2011.
IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp) and R (R Core
Team 2013. R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. URL http://www.R-project.org/) for data analysis.
Results are shown for the overall sample. Our initial approach would be to describe
the population stratified by sex, but after primary analysis, there was no relevant difference
in effect of ECG abnormalities in men when compared to women. Also, in the SCORE (8)
there is no risk interaction for the included covariates with sex.
RESULTS
The participants were followed for 8±1 years, with a resulting sample of 13,428
subjects (6039 men). The Table 1 shows the distribution of the
characteristics stratified by the presence of MEA. In general, those with the prevalent
abnormalities were older (57 versus 51 years), had higher prevalence of hypertension (59.1%
52
versus 32.1%), diabetes (32.6% versus 17.6%) and obesity (30.3% versus 21.6%).
Furthermore, the distribution of the population in the high-risk category did not differ
significantly between those with and without prevalent MEA (5.8% and 5.7% respectively).
Differences in smoking and dyslipidemia were also less relevant. We assessed education, as
a proxy for socioeconomic status. The participants with low education were more frequent
in the group with prevalent MEA than those without MEA (5.1% versus 10.1%). Those
without a college degree also had more MEA (59.9% versus 45.5%).
Table 2 demonstrates the prevalence of specific MEA and their related hazard ratios
(HR). Most frequent abnormalities were major isolated ST-T abnormalities (3.2%, HR=3.03,
95% CI 2.06-4.47), complete right bundle branch block (1.4%, HR=2.71 95% CI 1.59-4.66)
and major QT prolongation (0.9%, HR=2.92 95% CI 1.64-5.20) and the overall prevalence
of any MEA was 6.4% (HR=3.83, 95%CI 2.95-4.99). The absence of major Q wave
abnormalities is explained by the exclusion criteria of prevalent CVD - a major Q wave on
baseline ECG.
traditional cardiovascular risk factors are shown in Table 3. MEA remained a strong
predictor of mortality with a potential of more than doubling individual risk (HR 2.3 95%
CI 1.7-2.9). Cox regression stratified by sex can be seen in appendix (supplemental Table
1).
In the competitive risk analysis using the Fine and Gray model (table 3), the increase
in risk for cardiovascular mortality, after adjustment for the same parameters used for the
Cox model, was higher than the increase in risk for all-cause mortality, with a HR of 4.6
53
(95% CI 3.0-7.0). The cumulative risk curves comparing cardiovascular and non-
cardiovascular deaths are represented in Figure 3. It is evident that for the population with
prevalent MEA, risk of cardiovascular death becomes very similar to risk of death from all
other causes, although both outcomes have an increase in risk that can be attributable to the
presence of MEA. In the other hand, in the population without MEA the risk of non-
cardiovascular death is clearly much higher.
DISCUSSION
than doubling individual risk after multivariate analysis. When further exploring these
findings, the increase in risk was evident for both men and women and for low and high-risk
strata defined by the SCORE risk estimator. These findings corroborate with the idea that
there is little space for risk reclassification for low-risk individuals. In the cardiovascular
mortality analysis, additional risk of prevalent MEA became even more evident, doubling
the HR obtained in the Cox model for all-cause mortality. The competitive risk analysis also
demonstrated that non-cardiovascular death also increased in the population with prevalent
MEA (Figure 3).
Our study differs from the previously mentioned in some aspects. First, our cohort is
composed of relatively younger subjects (mean age of 52 years old), with a high prevalence
of diabetes and comprising a larger sample size (except for the studies conducted in the WHI
representing a relatively specific population).We are also the first to our knowledge to report
such findings in a Latin American setting. The Brazilian population is quite diverse and
information about Brazilian and Latin American population is lacking in the literature. We
also used the MC for describing the ECG alterations of interest. The MC has been the
standard tool in large population studies, providing opportunity for replication and perhaps
clinical use. For that reason, we also reported on the specific selected ECG abnormalities
(Table 2). ST-T abnormalities, complete right bundle branch block and major QT
prolongation were the most frequent and can be easily assessed clinically as can complete
55
left and indeterminate bundle branch blocks, which were also relatively frequent in our
study.
Another aspect in which our study differs from the others is that it reports only fatal
outcomes, which are less subject to bias then softer non-fatal events. It also reports curves
from the competitive risk model comparing cardiovascular and non-cardiovascular events.
Although the ECG is a tool typically used for cardiovascular assessment, it draws attention
to the fact it was able to predict all-cause mortality. Raghunath and colleagues (27) have also
recently explored the relationship between ECG and all-cause mortality. In a different
approach, using a deep neural network, they were able to build a risk prediction score for
all-cause mortality based on age, sex and ECG traces characteristics identified by the neural
network. The ROC curves obtained in the model were very satisfactory (area under the curve
of 0.876).
For purposes of risk classification, previous studies have also demonstrated that
conventional risk prediction tools can be improved by adding ECG information to a
prediction equation. Jørgensen et al incorporated ECG abnormalities defined by the MC into
the Framingham and the European SCORE risk score. This was done in participants older
than 65 years from the Copenhagen Heart Study. The presence of the defined
electrocardiographic abnormalities improved reclassification in the risk scores in 7.1% (95%
CI 6.7 - 9.0%) for fatal events and 4.2% (95% CI 3.5 - 5.6%) for non-fatal events. The new
proposed risk score was also internally validated by bootstrapping (28). On the other hand,
Groot et al demonstrated that the incorporation of ECG markers into the SCORE did not
result in significant reclassification of subjects. This study, conducted in a Dutch cohort,
followed a smaller number of participants (2,370) of younger age (mean age 48 years old)
and used different electrocardiographic variables (29).
In our paper, we opted to report the NRI, since it provides practical information as to
how patients could theoretically be reclassified after addition of a potential risk modifier (in
our case MEA). Our model provided 15% of correct reclassification in contrast to 3% of
misclassifications, with a net benefit for up reclassification of 12%. However, it is important
to state that, even though correct reclassifications in the event group were five times more
frequent than misclassifications in the non-event group (in relative terms), absolute numbers
of misclassifications were higher, since the non-event group was much more numerous.
Extrapolating a line of thought from randomized controlled trials in which magnitude of
benefit is measured with the number needed to treat, in our case, we would misclassify 4.3
56
individuals for each correct reclassification. Our model did not result in any down
reclassification for subjects without MEA. This can be explained by the fact the SCORE is
good enough to determine that being in a higher risk category is stronger to determine risk
than not having MEA. Figure 2 clearly illustrates this concept. The curves for survival
probability for intermediate risk participants without MEA still performed worse than any
curves for the low risk group. That means that not having MEA did not sufficiently reduce
only when MEA were present to affect reclassification purposes. Although there was no
relevant improvement in the C statistic, that usually is the case when evaluating effect of
incorporating a new variable to a preexisting model. This is actually one of the reasons why
a different and more practical approach in evaluating prediction models modifications such
as the NRI came into existence.
Compared to CACS, reclassification with the ECG, although not as accurate, was at
least comparable especially when considering its much lower cost, excellent safety profile
and its availability in primary care centers worldwide. In a primary prevention setting, taking
these facts into consideration and the favorable safety profile of statin use, the place of ECG
in risk stratification merits more attention. Its use might help decision-making and
individualization in eligibility for statin treatment, especially for the intermediate risk
patients.
Our study has some limitations. Firstly, we were not able to provide external
validation of our findings to better confirm the reclassification obtained. Another limitation
is that ELSA_Brasil is composed of only civil servants from six Brazilian state capitals, one
57
in the south, one in northeast and the other four in southeast. Thus, the generalizability our
findings to the Brazilian population should be done with caution.
Among the study's strengths, we can emphasize the large sample size and the detailed
information on participant's biological, socioeconomic and clinical data. Follow up of almost
ten years with very little loss and the standardization in ECG is also worth mentioning. Data
from LMIC, that faced a later epidemiological transition, when compared to high-income
countries, where most studies took place, also brings novelty to our findings. Finally, we
were able to explore our findings in multiple analyses, through Kaplan-Meier curves from
overall sample and risk strata, multivariate regression analysis, all cause, cardiovascular and
non-cardiovascular mortality and the potential in risk reclassification. All related to only
fatal outcomes. Altogether, this brings robustness to our findings and potential for clinical
applicability.
ACKNOLEDGEMNT
FUNDING
The ELSA Brasil baseline study was supported by the Brazilian Ministries of Health
supported by research grant from CNPq. A.L.P.R. is also a supported by a research grant
(Pesquisador Mineiro) from FAPEMIG, the research agency of the State of Minas Gerais,
Brazil.
DISCLOSURES
None
58
REFERENCES
1. Evaluation IoHMa. Findings from the Global Burden of Disease Study 2017. The
Lancet: IHME; 2018.
2. Nascimento BR BL, Moraes DN, Ribeiro AL. Global health and cardiovascular
disease. Heart. 2014 Nov;100(22):1743-9. PubMed PMID: 25327515.
3. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et
al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease.
Circulation. 2019 Mar 17:CIR0000000000000678. PubMed PMID: 30879355.
4. Cardiologia SBd. www.cardiometro.com.br 2019 [cited 2019 09/08/2019].
5. CDC. Preventable Deaths from Heart Disease & Stroke 2018. Available from:
https://www.cdc.gov/vitalsigns/.
6. Kaptoge S, Pennells L, De Bacquer D, Cooney MT, Kavousi M, Stevens G, et al.
World Health Organization cardiovascular disease risk charts: revised models to estimate
risk in 21 global regions. The Lancet Global Health. 2019;7(10):e1332-e45.
7. Yusuf S, Bosch J, Dagenais G, Zhu J, Xavier D, Liu L, et al. Cholesterol Lowering
in Intermediate-Risk Persons without Cardiovascular Disease. The New England journal of
medicine. 2016 May 26;374(21):2021-31. PubMed PMID: 27040132.
8. Cardiology ESo. SCORE Risk Charts - The European cardiovascular disease risk
assessment model escardio.org/Edication/Practice-Tools/CVD-prevention-
toolbox/SCORE-Risk-Charts2019 [cited 2019 12/08/2019].
9. Association ACoCaAH. 2013 ACC/AHA Guideline on the Assessment of
Cardiovascular Risk A Report of the American College of Cardiology/American Heart
Association Task Force on Practice Guidelines. Circulation. 2013.
10. Ridker PM. Comparing Cardiovascular Risk Prediction Scores. Annals of internal
medicine. 2015;162:313-4.
11. Lin JS, Evans CV, Johnson E, Redmond N, Coppola EL, Smith N. Nontraditional
Risk Factors in Cardiovascular Disease Risk Assessment: Updated Evidence Report and
Systematic Review for the US Preventive Services Task Force. JAMA. 2018 Jul
17;320(3):281-97. PubMed PMID: 29998301.
12. Rezaeian P, Ezekiel DN, Diaz T, Blumenthal R, Darwin C, McClelland R, et al.
Association of Major and Minor Electrocardiographic Abnormalities with Cardiovascular
Morbidity and Mortality in the Multi-Ethnic Study of Atherosclerosis Cohort. Journal of the
American College of Cardiology. 2016;67(13):831.
13. Shafiee G SF, Alizadeh M, Arzaghi SM, Maleki A, Esfahani MA, TajallizadeKhoob
Y, Larijani B, Fakhrzadeh H. Electrocardiogram abnormalities and risk of cardiovascular
mortality and all-cause mortality in old age: the Kahrizak Elderly Study (KES). Archives of
gerontology and geriatrics. 2014;59(1):69-73. PubMed PMID: 24794877.
14. Jonas DE, Reddy S, Middleton JC, Barclay C, Green J, Baker C, et al. Screening for
Cardiovascular Disease Risk With Resting or Exercise Electrocardiography: Evidence
Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2018 Jun
12;319(22):2315-28. PubMed PMID: 29896633.
15. Aquino EM, Barreto SM, Bensenor IM, Carvalho MS, Chor D, Duncan BB, et al.
Brazilian Longitudinal Study of Adult Health (ELSA-Brasil): objectives and design.
American journal of epidemiology. 2012 Feb 15;175(4):315-324. PubMed PMID:
22234482.
16. Force USPST, Bibbins-Domingo K, Grossman DC, Curry SJ, Davidson KW, Epling
JW, Jr., et al. Statin Use for the Primary Prevention of Cardiovascular Disease in Adults: US
Preventive Services Task Force Recommendation Statement. JAMA. 2016 Nov
15;316(19):1997-2007. PubMed PMID: 27838723.
59
17. James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J,
et al. 2014 evidence-based guideline for the management of high blood pressure in adults:
report from the panel members appointed to the Eighth Joint National Committee (JNC 8).
JAMA. 2014 Feb 5;311(5):507-20. PubMed PMID: 24352797.
18. Schmidt MI DB, Mill JG, Lotufo PA, Chor D, Barreto SM. Aquino EM, Passos VM,
Matos SM, Molina MC, Carvalho MS, Bensenor IM. Cohort Profile: Longitudinal Study of
Adult Health (ELSA-Brasil). International journal of epidemiology. 2015;44(1):68-75.
PubMed PMID: 24585730. Pubmed Central PMCID: 4339754.
19. Mozaffarian D BE, Go AS, Arnett DK, Blaha MJ, Cushman M, Das SR, de Ferranti
S, Despres JP, Fullerton HJ, Howard VJ, Huffman MD, Isasi CR, Jimenez MC, Judd SE,
Kissela BM, Lichtman JH, Lisabeth, LD, Liu S, Mackey RH, Magid DJ, McGuire DK,
Mohler ER 3rd, Moy CS, Muntner P, Mussolino ME, Nassir K, Neumar RW, Nichol G,
Palaniappan L, Pandey DK, Reeves MJ, Rodriguez CJ, Rosamond W, Sorlie PD, Stein J,
Towfighi A, Turan TN, Virani SS, Woo D, Yeh RW, Turner MB, American Heart
Association Statistics, Committee Stroke Statistics, Subcommittee. Heart Disease and Stroke
Statistics-2016 Update: A Report From the American Heart Association. Circulation.
2016;133(4):e38-e360. PubMed PMID: 26673558.
20. IBGE IBDGEE-. Características Étnico-raciais da População um estudo das
categorias de classificação de cor ou raça 2008. 2008.
21. Ribeiro ALP PS, Bergmann K, Ladeira RM, Oliveira RAM, Lotufo PA, Mill JG,
Barreto SM. Challenges to implementation of the ECG reading center in ELSA-Brazil.
Revista de Saúde Pública. 2013;47:87-94.
22. Macfarlane PW DB, Clark E. The University of Glasgow (Uni-G) ECG analysis
program. Computers in Cardiology. 2005;32:451-4.
23. Pinto-Filho MM, Brant LCC, Foppa M, Garcia-Silva KB, Mendes de Oliveira RA,
de Jesus Mendes da Fonseca M, et al. Major Electrocardiographic Abnormalities According
to the Minnesota Coding System Among Brazilian Adults (from the ELSA-Brasil Cohort
Study). The American journal of cardiology. 2017 Jun 15;119(12):2081-7. PubMed PMID:
28450038.
24. Prineas RJ CR, Zhang ZM. The Minnesota Code Manual of Electrocardiographic
Findings. Second ed. Springer L, editor. London: Springer; 2010.
25. Denes P. Major and Minor ECG Abnormalities in AsymptomaticWomen and Risk
of Cardiovascular Events and Mortality. Journal of the American Medical Association. 2007.
26. Badheka AO, Singh V, Patel NJ, Deshmukh A, Shah N, Chothani A, et al. QRS
duration on electrocardiography and cardiovascular mortality (from the National Health and
Nutrition Examination Survey-III). The American journal of cardiology. 2013 Sep
1;112(5):671-7. PubMed PMID: 23726176.
27. Raghunath S, Cerna AEU, Jing L, vanMaanen DP, Stough J, Hartzel DN, Leader
JB, Kirchner HL, Stumpe MC, Hafez A, Nemani A, Carbonati T, Johnson KW, Young K,
Good CW, Pfeifer JM, Patel AA, Delisle BP, Alsaid A, Beer D, Haggerty CM and Fornwalt
BK. Prediction of mortality from 12-lead electrocardiogram voltage data using a deep neural
network. Nature Medicine, 2020; s41591-020-0870-z
28. Jorgensen PG, Jensen JS, Marott JL, Jensen GB, Appleyard M, Mogelvang R.
Electrocardiographic changes improve risk prediction in asymptomatic persons age 65 years
or above without cardiovascular disease. Journal of the American College of Cardiology.
2014 Sep 2;64(9):898-906. PubMed PMID: 25169175.
29. Groot A, Bots ML, Rutten FH, den Ruijter HM, Numans ME, Vaartjes I.
Measurement of ECG abnormalities and cardiovascular risk classification: a cohort study of
primary care patients in the Netherlands. The British journal of general practice : the journal
60
FIGURE LEGENDS
Figure 1: Kaplan-Meier curves for cumulative survival from all-cause mortality according
to presence of major electrocardiographic abnormality at baseline, in the general population.
Figure 3: Fine and Gray model for competitive risk from non-cardiovascular and
cardiovascular death between those with and without major electrocardiographic
abnormalities at baseline.
61
TABLES
Table 1: Characteristics of the Brazilian Longitudinal Study of Adult Health participants
with valid electrocardiogram at baseline (overall and stratified by presence of major
cardiovascular abnormalities according to Minnesota coding criteria) (n= 13428). P are
represented only for comparisons between stratified groups.
Characteristics Overall (13428) No major ECG Major ECG P
abnormalities abnormalities value
(12566) (862)
Age (years) 51(45-58) 51(45-58) 57(50-64) <0.001
Men 45% 44.6% 50.7% <0.001
Age group (years)* <0.001
35-44 22.8% 23.7% 9.3%
45-54 39.9% 40.6% 29.7%
55-64 27.7% 27% 37.9%
65-74 9.6% 8.7% 23.1%
Race** 0.030
Black 16.7% 16.5% 19.4%
Mixed 29.4% 29.6% 26.0%
White 53.9% 53.9% 54.5%
Heart rate (bpm) 65(59-71) 65(59-71) 64(57-70) <0.001
Systolic blood 119(108-130) 118(108-129) 127(116-142) <0.001
pressure (mmHg)
Diastolic blood 75(68-83) 75(68-82) 78(71-86) <0.001
pressure (mmHg)
Body mass index 26.3(23.7-29.5) 26.2(23.6-29.4) 27.3(24.7-30.8) <0.001
(kg/m²)
Fasting glucose 105(98-113) 104(98-113) 109(101-122) <0.001
(mg/dl)
Low-density 129(108-152) 129(108-152) 130(109-153) 0.470
lipoprotein
cholesterol (mg/dl)
High-density 55(47-65) 55(47-65) 54(46-63) 0.018
lipoprotein
cholesterol (mg/dl)
Total Cholesterol 213(187-240) 212(187-240) 214(190-242) 0.089
(mg/dl)
Risk Category* 0.874
Low 62.2% 62.3% 61.4%
Medium 32.1% 32% 32.8%
High 5.7% 5.7% 5.8%
Hypertension 33.8% 32.1% 59.1% <0.001
Diabetes Mellitus 18.6% 17.6% 32.6% <0.001
Dyslipidemia 48.7% 48.3% 54.5% <0.001
Obesity 22.1% 21.6% 30.3% <0.001
Present smoking 13.1% 13.1% 13.7% 0.606
Education
Low 5.4% 5.1% 10.1% <0.001
Less than superior 46.1% 45.5% 59.9% <0.001
bpm = beats per minute. Values are frequencies or medians and interquartile range. * For age group and SCORE risk score
category, P is represented for linear trend **Those who did not declare race were excluded from race analysis. Less frequent
race categories were also not analyzed (Asian and indigenous). Risk categorization was done according to the Systematic
Coronary Risk Evaluation (SCORE).
62
Table 2: Prevalence of major cardiovascular abnormalities and their respective crude hazard
ratios and 95% confidence interval, for the outcome of total mortality (n= 13428).
Table 3: Total and cardiovascular mortality rates and hazard ratio according to the presence of major
electrocardiographic abnormalities and multivariate models.
Overall No ECG ECG P Hazard ratio (95% CI)*
abn. abn.
M1 M2 M3 M4
Total 2.8% 2.5% 8.7% <0.001 3.8(2.9-5.0) 2.5(1.9-3.3) 2.4(1.8-3.1) 2.3(1.7-2.9)
mortality
Overall No ECG ECG P Relative Risk (95% CI)
abn. abn.
Multivariate Fine and Gray model**
CV 1.2% 1.1% 1.6% 0.193 4.6(3.0-7.0)
mortality
*model 1 was unadjusted, model 2 adjusted for sex, age group, race, model 3 for sex, age group, race and school years, model 4 for sex, age group,
race, school years and cardiovascular risk factors (smoking, hyperlipidemia, hypertension, obesity and diabetes). **Model adjusted for sex, age
group, race, school years and cardiovascular risk factors (smoking, hyperlipidemia, hypertension, obesity and diabetes). abn=abnormality;
CV=cardiovascular; CI=confidence interval.
63
Table 4: Systematic Coronary Risk Evaluation score reclassification for the outcome of
FIGURES
Figure 1:
65
Figure 2:
66
Figure 3:
Left panel represents probability of event (cardiovascular death in red lines and non-cardiovascular death in
blue line) in the populations free of major electrocardiographic abnormalities (MEA). The right panel
represents the same events in those with MEA present.
67
SUPLEMENTAL MATERIAL
TABLES
Table 1: Total mortality rates according to sex and presence of major electrocardiographic abnormalities
and the hazard ratio of the presence of major electrocardiographic abnormality in a multivariate Cox
regression model.
Men (6041)
Overall No ECG ECG P Hazard ratio (95% CI)*
abn. abn.
M1 M2 M3 M4
Mortality 3.8% 3.2% 11.2% <0.001 3.6(2.6-5.1) 2.5(1.8-3.5) 2.4(1.7-3.4) 2.2(1.6-3.1)
Women (7396)
Overall No ECG ECG P Hazard ratio (95% CI)*
abn. abn.
M1 M2 M3 M4
Mortality 2.1% 1.8% 6.1% <0.001 3.8(2.5-5.9) 2.6(1.7-4.0) 2.4(1.5-3.8) 2.3(1.5-3.6)
*model 1 was unadjusted, model 2 adjusted for age group + race, model 3 for age group, race and school years, model 4 for age group, race,
school years and cardiovascular risk factors (smoking, hyperlipidemia, hypertension, obesity and diabetes). abn=abnormality; CI=confidence
interval
68
FIGURES
Figure 2: Receiver operating characteristic curves for the SCORE risk calculator for
cardiovascular mortality and for the modified risk score including major
electrocardiographic abnormalities.
70
Marcelo M. Pinto-Filho, MD, MSca,b, Luisa C. C. Brant, MD, PhDa,b, Murilo Foppa, MD,
PhDc,d, Scheine L Canhada, MScd, Paulo A. Lotufo, MD, PhDe,f, Renan B Pereirad Sandhi
M. Barreto, MD, PhDa, Peter W. Macfarlane, DSc, FRCPg, Antonio L. P. Ribeiro, MD,
PhDa,b
Hospital Borges da Costa. Alfredo Balena Av, 110 Funcionários ZIP 30130-100 Belo
ABSTRACT
METHODS: Participants free of cardiovascular disease, without prevalent MEA from the
Longitudinal Study of Adult Health (ELSA-Brasil) were included. A 12-lead
electrocardiogram was obtained from the participants both at baseline (2008-2009) and in a
subsequent assessment (2012-2013), and coded according to the MC. To evaluate potential
predictors of incident MEA, we compared participant
cardiovascular risk factors such as hypertension, obesity, dyslipidemia, diabetes and age by
their medians and frequencies.
RESULTS: After exclusion criteria, 11608 participants (6433 women) were suitable for
analysis. The clinical characteristics were comparable between the participants with and
without MEA in the follow-up, for both men and women. Overall incidence of new MEA
were 10.9% for men and 11.1% for women. Among specific MEA, the highest incidence
was: for isolated major ST-T abnormalities (4.3% in men, 3.9% in women), major Q waves
(2.8% in both men and women) and complete right bundle branch block (2.0% in men and
2.4% in women). Body mass index (BMI) was the only clinical characteristic found to be
associated with incident MEA, although in small magnitude and in opposite direction for
men and women.
CONCLUSIONS: Among participants without CVD and MEA at baseline of the ELSA-
Brasil, the incidence of MEA was over 10% in a mean four-year follow up. None of the
potential clinical predictors evaluated were consistently correlated to incident MEA.
Keywords:
INTRODUCTION
Cardiovascular diseases (CVD) treatment, prevention and awareness have been
improving constantly since the 20th century, leading to an observed decline in CVD
mortality worldwide in the 21st century (1, 2). Nevertheless, CVD remains the main cause
of death both in high- and low and middle-income countries countries. Among non-
communicable diseases, CVD has great potential for prevention (1). It is noteworthy that
CVD mortality has decreased less steeply in low and middle- income countries (3, 4). Better
understanding of disease dynamics and more accurate risk prediction may affect how we
treat and prevent CVD. The electrocardiogram is a well-recognized tool for assessment of
numerous cardiovascular conditions. Electrocardiographic abnormalities have been shown
to be associated with poor cardiovascular outcomes (5-9), being considered a marker of
subclinical CVD (10, 11). There is a paucity of information in the literature regarding
incidence of electrocardiographic abnormalities, with no information reported for the
Brazilian population. The Longitudinal Study of Adult Health (ELSA-Brasil) is an ongoing
cohort study conducted in Brazil, which aims to contribute to better understanding of CVD
and CVD determinants, with the specificity of being conducted in a middle-income country.
In the present study, we aim to determine the incidence of new major electrocardiographic
abnormalities (MEA) in the participants of the ELSA-Brasil cohort, as well as to evaluate
ld be associated with the incident abnormalities.
Analyses of the baseline electrocardiograms (ECG) of ELSA-Brasil cohort participants have
been subject to previous publications (12, 13).
METHODS
Study Design and Participants
This is a prospective analysis from the ELSA-Brasil cohort, a longitudinal study that
recruited 15105 civil servants of both sexes, aged 35 to 74 years old, living in six Brazilian
state capitals. In addition to a 12-lead ECG, sociodemographic, anthropometric and clinical
data were collected from all participants at baseline. The detailed methodology has been
published elsewhere (14). For the present analyses, we selected participants without known
cardiovascular disease (defined by self-report of heart failure, myocardial infarction, stroke,
previous myocardial revascularization or a major Q wave in baseline ECG) and without
major electrocardiographic abnormalities at baseline. Participants that were lost to follow
excluded, resulting in a final sample of 11608 participants. Baseline data were collected at
73
2008 and, at a mean four-year interval, a new round of data collection took place
(2012/2013), including the obtainment of new ECG. The ELSA-Brasil study was approved
by local ethics committees of the six institutions where participants were recruited.
ECG record
MEA can be defined by (17): major Q waves (old myocardial infarctions, MC 1-1
and 1-2), minor Q waves plus ST-T abnormalities (possible old myocardial infarction MC 1-
3 plus MC 4-1 or 4-2 or 5-1 or 5-2), major isolated ST-T abnormalities (MC 4-1 or 4-2 or 5-
1 or 5-2), left ventricular hypertrophy plus ST-T abnormalities (MC 3-1 plus MC 4-1 or 4-2
or 5-1 or 5-2), intraventricular conduction abnormalities (complete/intermittent right and left
bundle branch block, non-specific intraventricular block, MC 7-1 or 7-2 or 7-4), right bundle
branch block plus left anterior divisional block (MC 7-8), major QT prolongation index (QTi
atrial fibrillation/flutter (MC 8-3), supraventricular tachycardia (MC 8-4-2),
atrioventricular (AV) conduction abnormalities (second and third degree AV block, artificial
pacemaker, ventricular pre-excitation such as Wolff Parkinson White syndrome, MC 6-1 or
6-2 or 6-4 or 6-8).
Study Variables
and their relation to new MEA. Hypertension was defined by its classical definition, as a
0 mmHg
(20, 21), or by self-reported use of anti-hypertensive medication. Participants were
considered to have diabetes if they were receiving treatment for this condition or if they had
a known diagnosis of diabetes or if their fasting glucose were , post prandial
, (22). For the present study, presence of
dyslipidemia was defined
74
-lowering medication
acknowledge that there is variation in the definition of dyslipidemia among societies and
thresholds vary through time and evidence progression, but we feel that, for classification
purposes this was a reasonable and traditional cut off (23). Obesity was defined as a body
smokers were those with current use of tobacco. Cardiovascular
risk was calculated by the pooled cohort equations adjusted by our background and race
pattern (24). The risk categor
Heart Association guideline publication, (25) and has practical purposes, since there still
remains clinical doubt on best management of subjects with risk between 5 and 7.5% (26).
Race was self-reported according to the Brazilian National Census (27)
schooling years were used as a proxy for socioeconomic status. Those with less then the first
four years of education completed were considered to have low educational attainment.
Statistical Analyses
RESULTS
Among the 15105 participants recruited for the ELSA-Brasil study, the present
analysis included, after pre-specified exclusion criteria 11609 participants. Due to death or
loss to follow up 1091 participants did not participate in the second round of assessment and
144 ECG from those obtained were missing. From baseline, 1277 participants had MEA and
124 ECG were missing. In respect to data from self-reported cardiovascular disease, 399
participants had missing data at baseline and 461 had self-reported cardiovascular disease at
baseline. Exclusion flow is detailed in figure 1.
75
Among the selected participants, when stratified by incident MEA identified at the
second round of examinations, there was no significant difference between groups in almost
none of the clinical characteristics. Exception can be made for a statistically significant
difference in the BMI that was lower for men with incident MEA (26.4 kg/m² versus 26.0
kg/m²) and higher for women (25.9 kg/m² versus 26.6 kg/m²). A difference in obesity rate was
found only for the women. Those with MEA were more obese than those without MEA
(27.2% versus 22.8%). The clinical characteristics of the study population can be found in
table 1. The differences in BMI were small in clinical relevance and happened in opposite
directions for men and women, which limits its interpretation as a true association with
incident MEA. Obesity was also considered to be a weak association since its data derive
from BMI and happened only for the women. Hence, strength of association was not further
explored.
The specific MEA are listed in table 2. Among these participants without MEA at
baseline and free of CVD, major isolated ST-T abnormalities were the most common novel
abnormalities (4.3% in men and 3.9% in women), followed by major Q-wave (2.8% in both
men and women) and complete right bundle branch block (2.0% in men and 2.4% in
women). In our sample the development of new MEA was as frequent in men as in women
(10.9% and 11.1%, respectively).
DISCUSSION
In the ELSA-Brasil cohort, 11% of the participants developed new MEA in an mean
follow up of four years. Although this rate is fairly high, there was no difference in the rate
of MEA according to sex or most clinical characteristics. We found BMI to be associated
with prevalent MEA for both sexes, though the direction of the association was opposite and
the numerical difference small, making type I error a possible explanation for the finding.
Higher prevalence of obesity for women in participants with incident MEA was also found.
For the same reasons as the ones mentioned for BMI, its significance remains questionable.
Further exploration of this data, with extended follow up and, consequently, more events,
might help us clarify if there is a true association between obesity or BMI and the incidence
of MEA.
There are some potential explanations for our negative findings. First, our exclusion
criteria left us with a lower risk population. In previous publication, regarding baseline
electrocardiographic data from the ELSA-Brasil population (13) some findings elucidate this
76
assumption. It was demonstrated that MEA were more common in men then in women, and
that MEA prevalence increased as cardiovascular risk factors increased. By excluding these
participants (more men and more high risk), the sex effect could have disappeared and the
risk profile of the analyzed population, reduced. Indeed, our present sample had a median
cardiovascular risk of 2.6% when computed altogether, 5,1% for men, 1,6% for women
using . Even among men and women who developed MEA,
mean cardiovascular risk was under 7.5%, which is considered to be less then intermediate
risk by current literature (26). That might have led us with fewer events and therefore, lower
power.
Another potential explanation for our findings is the relatively short follow up. A
mean of four years of follow-up might have been too little time for the development of new
MEA. Soliman et al demonstrated the association between type 1 diabetes and new
electrocardiographic abnormalities in a primary prevention cohort of diabetic patients (a high
risk condition per se). Mean age was lower in that cohort (35 years old) but follow up was
through sixteen years, with a 13.1% incidence of new MEA(10). The authors demonstrated
that for every 10% increase in glycosylated hemoglobin there was an increase of 16% in risk
for development of new MEA.
One relevant limitation in our study is the aforementioned short period of follow- up.
Nevertheless, we expect to evaluate data from eight years of follow up in a short coming
future. That might provide us more power to identify potential predictors of MEA. As
noticeable strengths, the standardization in ECG obtainment, the detailed clinical and social
data available, provided us opportunity for numerous assessments. Most importantly, this is
also the first study to report incidence data on electrocardiographic abnormalities in the
Brazilian population and, to our knowledge, it is the first to report incidence of MEA,
standardized by the MC in a general population. Our study will also give us the opportunity
for future correlations of specific MEA, such as major Q waves and incident cardiovascular
disease such as myocardial infarction, once data on cardiovascular outcomes are available.
In conclusion, incidence of MEA were high in the studied population, but we could
not identify, in this time-period, any predictors for the development of new MEA. They were
as common in men as in women and there was no effect of race, educational level or
traditional cardiovascular risk factors in the incidence of MEA.
77
ACKNOLEDGEMNTS
FUNDING
The ELSA Brasil baseline study was supported by the Brazilian Ministries of Health
supported by research grant from CNPq. A.L.P.R. is also a supported by a research grant
(Pesquisador Mineiro) from FAPEMIG, the research agency of the State of Minas Gerais,
Brazil.
DISCLOSURES
None
REFERENCES
1. Benjamin EJ, Muntner P, Alonso A, Bittencourt MS, Callaway CW, Carson AP, et
al. Heart Disease and Stroke Statistics-2019 Update: A Report From the American Heart
Association. Circulation. 2019 Mar 5;139(10):e56-e528. PubMed PMID: 30700139.
2. Association AH. https://healthmetrics.heart.org/total-cardiovascular-disease-
mortality-in-the-united-states-1999-2016/ American Heart Association web page: American
Heart Association; 2019 [updated JUNE 26, 2018; cited 2019 11/06/2019].
3. Ribeiro AL, Duncan BB, Brant LC, Lotufo PA, Mill JG, Barreto SM. Cardiovascular
Health in Brazil: Trends and Perspectives. Circulation. 2016 Jan 26;133(4):422-33. PubMed
PMID: 26811272.
4. Roth GA, Huffman MD, Moran AE, Feigin V, Mensah GA, Naghavi M, et al. Global
and regional patterns in cardiovascular mortality from 1990 to 2013. Circulation. 2015 Oct
27;132(17):1667-78. PubMed PMID: 26503749.
5. Rezaeian P, Ezekiel DN, Diaz T, Blumenthal R, Darwin C, McClelland R, et al.
Association of Major and Minor Electrocardiographic Abnormalities with Cardiovascular
Morbidity and Mortality in the Multi-Ethnic Study of Atherosclerosis Cohort. Journal of the
American College of Cardiology. 2016;67(13):831.
6. Shafiee G SF, Alizadeh M, Arzaghi SM, Maleki A, Esfahani MA, TajallizadeKhoob
Y, Larijani B, Fakhrzadeh H. Electrocardiogram abnormalities and risk of cardiovascular
mortality and all-cause mortality in old age: the Kahrizak Elderly Study (KES). Archives of
gerontology and geriatrics. 2014;59(1):69-73. PubMed PMID: 24794877.
7. Gencer B, Butler J, Bauer DC, Auer R, Kalogeropoulos A, Marques-Vidal P, et al.
Association of electrocardiogram abnormalities and incident heart failure events. American
heart journal. 2014 Jun;167(6):869-75 e3. PubMed PMID: 24890537. Pubmed Central
PMCID: 4081858.
78
8. Rautaharju PM, Zhang ZM WJ, Gregg RE, Haisty WK, Kucharska-Newton AM,
Rosamond WD, Soliman EZ. Electrocardiographic predictors of coronary heart disease and
sudden cardiac deaths in men and women free from cardiovascular disease in the
Atherosclerosis Risk in Communities study. Journal of the American Heart Association.
2013;2(3):e000061. PubMed PMID: 23723252. Pubmed Central PMCID: 3698763.
9. Zhang Z-m, Prineas RJ, Case D, Psaty BM, Suzuki T, Burke GL. Gender Differences
Between the Minnesota Code and Novacode Electrocardiographic Prognostication of
Coronary Heart Disease in the Cardiovascular Health Study. The American journal of
cardiology. 2011;107(6):817-20.e1.
10. Soliman EZ, Backlund JY, Bebu I, Li Y, Zhang ZM, Cleary PA, et al. Progression
of Electrocardiographic Abnormalities in Type 1 Diabetes During 16 Years of Follow-up:
The Epidemiology of Diabetes Interventions and Complications (EDIC) Study. Journal of
the American Heart Association. 2016 Mar 14;5(3):e002882. PubMed PMID: 26976878.
Pubmed Central PMCID: 4943265.
11. Denes P LJ, Lloyd Jones DM, Prineas RJ, Greenland P. Major and Minor ECG
Abnormalities in AsymptomaticWomen and Risk of Cardiovascular Events and Mortality.
Journal of the American Medical Association. 2007;297(9):978-85.
12. Pinto MMF, Brant LCC, Padilha-da-Silva JL, Foppa M, Lotufo PA, Mill JG, et al.
Electrocardiographic Findings in Brazilian Adults without Heart Disease: ELSA-Brasil. Arq
Bras Cardiol. 2017 Nov;109(5):416-24. PubMed PMID: 28977056. Pubmed Central
PMCID: 5729777.
13. Pinto-Filho MM, Brant LCC, Foppa M, Garcia-Silva KB, Mendes de Oliveira RA,
de Jesus Mendes da Fonseca M, et al. Major Electrocardiographic Abnormalities According
to the Minnesota Coding System Among Brazilian Adults (from the ELSA-Brasil Cohort
Study). The American journal of cardiology. 2017 Jun 15;119(12):2081-7. PubMed PMID:
28450038.
14. Aquino EM, Barreto SM, Bensenor IM, Carvalho MS, Chor D, Duncan BB, et al.
Brazilian Longitudinal Study of Adult Health (ELSA-Brasil): objectives and design.
American journal of epidemiology. 2012 Feb 15;175(4):315-24. PubMed PMID: 22234482.
15. Ribeiro ALP PS, Bergmann K, Ladeira RM, Oliveira RAM, Lotufo PA, Mill JG,
Barreto SM. Challenges to implementation of the ECG reading center in ELSA-Brazil.
Revista de Saúde Pública. 2013;47:87-94.
16. Macfarlane PW DB, Clark E. The University of Glasgow (Uni-G) ECG analysis
program. Computers in Cardiology. 2005;32:451-4.
17. Prineas RJ CR, Zhang ZM. The Minnesota Code Manual of Electrocardiographic
Findings. Second ed. Springer L, editor. London: Springer; 2010.
18. Auer R BDC, Marques-Vidal P, Butler J, in L. J, Cornuz J, Satterfield R, Newman
A. B, Vittinghoff E, Rodondi N. Association ofMajor and Minor ECG
AbnormalitiesWithCoronaryHeartDiseaseEvents. Journal of the American Medical
Association. 2012;307(14):1497-505.
19. Kors AJ CR, Hannan PJ, Rautaharju PM, Folsom AR. Comparison of Computer-
assigned Minnesota Codes with the Visual Standard Method for New Coronary Heart
Disease Events. American journal of epidemiology. 2000;Vol. 151(8):790-7.
20. James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J,
et al. 2014 evidence-based guideline for the management of high blood pressure in adults:
report from the panel members appointed to the Eighth Joint National Committee (JNC 8).
JAMA. 2014 Feb 5;311(5):507-20. PubMed PMID: 24352797.
21. Cardiologia SBd. 7ª Diretriz Brasileira de Hipertensão Arterial. Arquivoas
Brasileiros de Cardiologia - -0066-782X.
2016;107(3):Supl 3.
79
22. Schmidt MI DB, Mill JG, Lotufo PA, Chor D, Barreto SM. Aquino EM, Passos VM,
Matos SM, Molina MC, Carvalho MS, Bensenor IM. Cohort Profile: Longitudinal Study of
Adult Health (ELSA-Brasil). International journal of epidemiology. 2015;44(1):68-75.
PubMed PMID: 24585730. Pubmed Central PMCID: 4339754.
23. Mozaffarian D BE, Go AS, Arnett DK, Blaha MJ, Cushman M, Das SR, de Ferranti
S, Despres JP, Fullerton HJ, Howard VJ, Huffman MD, Isasi CR, Jimenez MC, Judd SE,
Kissela BM, Lichtman JH, Lisabeth, LD, Liu S, Mackey RH, Magid DJ, McGuire DK,
Mohler ER 3rd, Moy CS, Muntner P, Mussolino ME, Nassir K, Neumar RW, Nichol G,
Palaniappan L, Pandey DK, Reeves MJ, Rodriguez CJ, Rosamond W, Sorlie PD, Stein J,
Towfighi A, Turan TN, Virani SS, Woo D, Yeh RW, Turner MB, American Heart
Association Statistics, Committee Stroke Statistics, Subcommittee. Heart Disease and Stroke
Statistics-2016 Update: A Report From the American Heart Association. Circulation.
2016;133(4):e38-e360. PubMed PMID: 26673558.
24. Kaptoge S, Pennells L, De Bacquer D, Cooney MT, Kavousi M, Stevens G, et al.
World Health Organization cardiovascular disease risk charts: revised models to estimate
risk in 21 global regions. The Lancet Global Health. 2019;7(10):e1332-e45.
25. Association ACoCaAH. 2013 ACC/AHA Guideline on the Assessment of
Cardiovascular Risk A Report of the American College of Cardiology/American Heart
Association Task Force on Practice Guidelines. Circulation. 2014.
26. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et
al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease.
Circulation. 2019 Mar 17:CIR0000000000000678. PubMed PMID: 30879355.
27. IBGE IBDGEE-. Características Étnico-raciais da População um estudo das
categorias de classificação de cor ou raça 2008. 2008.
80
FIGURE
Death or lost to
follow-up: 1091
participants
TABLES
7. CONSIDERAÇÕES GERAIS
8. CONCLUSÕES
Em uma coorte de homens e mulheres de seis estados brasileiros, com idade média de
(51 anos), livres de doença cardiovascular na linha de base, a presença de AEM evidenciadas
pelo ECG obtido na linha de base do estudo, aumentou o risco de morte (total e
cardiovascular) em modelo multivariado, levando-se em conta sexo, idade, raça,
escolaridade e fatores de risco tradicionais para DCV. O risco adicional imposto por essas
alterações foi de relevante magnitude e é fundamental que estendamos a discussão do papel
do ECG na estratificação de risco cardiovascular dada a segurança do método e seu baixo
custo.
Ao ser incorporado a um modelo de predição de risco CV tradicional, (SCORE), a
presença de AEM proporcionou um índice de reclassificação líquida positivo. Dentre os
participantes que experienciaram o desfecho de morte cardiovascular, 15% foram
reclassificados corretamente ao contrário de apenas 3% de reclassificações incorretas dentre
os que não sofreram o evento.
A incidência de AEM nessa população também foi alta, mas não encontramos
incidência de AEM.
9. REFERÊNCIAS BIBLIOGRÁFICAS
1. Evaluation IoHMa. Findings from the Global Burden of Disease Study 2017. The
Lancet: IHME; 2018.
2. Nascimento BR BL, Moraes DN, Ribeiro AL. Global health and cardiovascular
disease. Heart. 2014 Nov;100(22):1743-9. PubMed PMID: 25327515.
3. JMA S. Transição epidemiológica e o estudo de carga de doença no Brasil. Ciência
e Saúde Coletiva. 2004;9(4):897-908.
4. Organization WH. 2008-2013 Action Plan for the Global Strategy for the Prevention
and Control of Noncommunicable Diseases. 2008.
5. Mozaffarian D BE, Go AS, Arnett DK, Blaha MJ, Cushman M, Das SR, de Ferranti
S, Despres JP, Fullerton HJ, Howard VJ, Huffman MD, Isasi CR, Jimenez MC, Judd SE,
Kissela BM, Lichtman JH, Lisabeth, LD, Liu S, Mackey RH, Magid DJ, McGuire DK,
Mohler ER 3rd, Moy CS, Muntner P, Mussolino ME, Nassir K, Neumar RW, Nichol G,
Palaniappan L, Pandey DK, Reeves MJ, Rodriguez CJ, Rosamond W, Sorlie PD, Stein J,
Towfighi A, Turan TN, Virani SS, Woo D, Yeh RW, Turner MB, American Heart
Association Statistics, Committee Stroke Statistics, Subcommittee. Heart Disease and Stroke
Statistics-2016 Update: A Report From the American Heart Association. Circulation.
2016;133(4):e38-e360. PubMed PMID: 26673558.
6. Ford ES, Greenlund KJ, Hong Y. Ideal cardiovascular health and mortality from all
causes and diseases of the circulatory system among adults in the United States. Circulation.
2012 Feb 28;125(8):987-95. PubMed PMID: 22291126. Pubmed Central PMCID: 4556343.
85
7. CDC. Preventable Deaths from Heart Disease & Stroke 2018. Available from:
https://www.cdc.gov/vitalsigns/.
8. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et
al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease.
Circulation. 2019 Mar 17:CIR0000000000000678. PubMed PMID: 30879355.
9. Piepoli MF, Hoes AW, Agewall S, Albus C, Brotons C, Catapano AL, et al. 2016
European Guidelines on cardiovascular disease prevention in clinical practice: The Sixth
Joint Task Force of the European Society of Cardiology and Other Societies on
Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of 10
societies and by invited experts)Developed with the special contribution of the European
Association for Cardiovascular Prevention & Rehabilitation (EACPR). European heart
journal. 2016 Aug 1;37(29):2315-81. PubMed PMID: 27222591. Pubmed Central PMCID:
4986030.
10. ACC/AHA Risk Calculator. Available from: http://www.cvriskcalculator.com/.
11. Framingham Heart Study. Available from:
https://www.framinghamheartstudy.org/fhs-risk-functions/cardiovascular-disease-10-year-
risk/.
12. Cardiology ESo. SCORE Risk Charts - The European cardiovascular disease risk
assessment model escardio.org/Edication/Practice-Tools/CVD-prevention-
toolbox/SCORE-Risk-Charts2019 [cited 2019 12/08/2019].
13. Yusuf S, Bosch J, Dagenais G, Zhu J, Xavier D, Liu L, et al. Cholesterol Lowering
in Intermediate-Risk Persons without Cardiovascular Disease. The New England journal of
medicine. 2016 May 26;374(21):2021-31. PubMed PMID: 27040132.
14. DeFilippis AP, Young R, Carrubba CJ, McEvoy JW, Budoff MJ, Blumenthal RS, et
al. An analysis of calibration and discrimination among multiple cardiovascular risk scores
in a modern multiethnic cohort. Annals of internal medicine. 2015 Feb 17;162(4):266-75.
PubMed PMID: 25686167. Pubmed Central PMCID: 4414494.
15. Ridker PM. Comparing Cardiovascular Risk Prediction Scores. Annals of internal
medicine. 2015;162:313-4.
16. Lin JS, Evans CV, Johnson E, Redmond N, Coppola EL, Smith N. Nontraditional
Risk Factors in Cardiovascular Disease Risk Assessment: Updated Evidence Report and
Systematic Review for the US Preventive Services Task Force. JAMA. 2018 Jul
17;320(3):281-97. PubMed PMID: 29998301.
17. Rezaeian P, Ezekiel DN, Diaz T, Blumenthal R, Darwin C, McClelland R, et al.
Association of Major and Minor Electrocardiographic Abnormalities with Cardiovascular
Morbidity and Mortality in the Multi-Ethnic Study of Atherosclerosis Cohort. Journal of the
American College of Cardiology. 2016;67(13):831.
18. Shafiee G SF, Alizadeh M, Arzaghi SM, Maleki A, Esfahani MA, TajallizadeKhoob
Y, Larijani B, Fakhrzadeh H. Electrocardiogram abnormalities and risk of cardiovascular
mortality and all-cause mortality in old age: the Kahrizak Elderly Study (KES). Archives of
gerontology and geriatrics. 2014;59(1):69-73. PubMed PMID: 24794877.
19. Gencer B, Butler J, Bauer DC, Auer R, Kalogeropoulos A, Marques-Vidal P, et al.
Association of electrocardiogram abnormalities and incident heart failure events. American
heart journal. 2014 Jun;167(6):869-75 e3. PubMed PMID: 24890537. Pubmed Central
PMCID: 4081858.
20. Rautaharju PM, Zhang ZM, Warren J, Gregg RE, Haisty WK, Kucharska-Newton
AM, et al. Electrocardiographic predictors of coronary heart disease and sudden cardiac
deaths in men and women free from cardiovascular disease in the Atherosclerosis Risk in
Communities study. Journal of the American Heart Association. 2013 May 30;2(3):e000061.
PubMed PMID: 23723252. Pubmed Central PMCID: 3698763.
86
21. Badheka AO, Singh V, Patel NJ, Deshmukh A, Shah N, Chothani A, et al. QRS
duration on electrocardiography and cardiovascular mortality (from the National Health and
Nutrition Examination Survey-III). The American journal of cardiology. 2013 Sep
1;112(5):671-7. PubMed PMID: 23726176.
22. Zhang Z-m, Prineas RJ, Case D, Psaty BM, Suzuki T, Burke GL. Gender Differences
Between the Minnesota Code and Novacode Electrocardiographic Prognostication of
Coronary Heart Disease in the Cardiovascular Health Study. The American journal of
cardiology. 2011;107(6):817-20.e1.
23. Rautaharju PM, Kooperberg C, Larson JC, LaCroix A. Electrocardiographic
abnormalities that predict coronary heart disease events and mortality in postmenopausal
women: the Women's Health Initiative. Circulation. 2006 Jan 31;113(4):473-80. PubMed
PMID: 16449726.
24. Machado DB, Crow RS, Boland LL, Hannan PJ, Taylor HA, Jr., Folsom AR.
Electrocardiographic findings and incident coronary heart disease among participants in the
Atherosclerosis Risk in Communities (ARIC) study. The American journal of cardiology.
2006 Apr 15;97(8):1176-81. PubMed PMID: 16616022.
25. Triola B, Olson MB, Reis SE, Rautaharju P, Merz CNB, Kelsey SF, et al.
Electrocardiographic Predictors of Cardiovascular Outcome in Women. Journal of the
American College of Cardiology. 2005;46(1):51-6.
26. Jonas DE, Reddy S, Middleton JC, Barclay C, Green J, Baker C, et al. Screening for
Cardiovascular Disease Risk With Resting or Exercise Electrocardiography: Evidence
Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2018 Jun
12;319(22):2315-28. PubMed PMID: 29896633.
27. Chou R and High Value Care Task Force of the American College of P. Cardiac
screening with electrocardiography, stress echocardiography, or myocardial perfusion
imaging: advice for high-value care from the American College of Physicians. Annals of
internal medicine. 2015;162(6):438-47. PubMed PMID: 25775317.
28. Aquino EM, Barreto SM, Bensenor IM, Carvalho MS, Chor D, Duncan BB, et al.
Brazilian Longitudinal Study of Adult Health (ELSA-Brasil): objectives and design.
American journal of epidemiology. 2012 Feb 15;175(4):315-24. PubMed PMID: 22234482.
29. Lotufo PA. [Setting up the longitudinal study for adult health (ELSA-Brasil]. Rev
Saude Publica. 2013 Jun;47 Suppl 2:3-9. PubMed PMID: 24346715. Construcao do Estudo
Longitudinal de Saude do Adulto (ELSA-Brasil).
30. Barreto SM, Ladeira RM, Bastos MdSCBd, Diniz MdFHS, Jesus EAd, Kelles SMB,
et al. Estrategias de identificacao, investigacao e classificacao de desfechos incidentes no
ELSA-Brasil. Revista de Saúde Pública. 2013;47:79-86.
31. Mill JG PK, Griep RH, Goulart A, Foppa M, Lotufo PA, Maestri MK, Ribeiro AL,
Andreao RV, Dantas EM, Oliveira I, Fuchs SC, Cunha RS, Bensenor IM. Afericoes e exames
clinicos realizados nos participantes do ELSA-Brasil. Revista de Saúde Pública. 2013;47:54-
62.
32. Aquino EM BS, Bensenor IM, Carvalho MS, Chor D, Duncan BB, Lotufo PA, Mill
JG, Molina Mdel C, Mota EL, Passos VM, Schmidt MI, Szklo M. Brazilian Longitudinal
Study of Adult Health (ELSA-Brasil): objectives and design. American journal of
epidemiology. 2012;175(4):315-24. PubMed PMID: 22234482.
33. Ribeiro ALP PS, Bergmann K, Ladeira RM, Oliveira RAM, Lotufo PA, Mill JG,
Barreto SM. Challenges to implementation of the ECG reading center in ELSA-Brazil.
Revista de Saúde Pública. 2013;47:87-94.
34. Pinto-Filho MM, Brant LCC, Foppa M, Garcia-Silva KB, Mendes de Oliveira RA,
de Jesus Mendes da Fonseca M, et al. Major Electrocardiographic Abnormalities According
to the Minnesota Coding System Among Brazilian Adults (from the ELSA-Brasil Cohort
87
Study). The American journal of cardiology. 2017 Jun 15;119(12):2081-7. PubMed PMID:
28450038.
35. Pinto MMF, Brant LCC, Padilha-da-Silva JL, Foppa M, Lotufo PA, Mill JG, et al.
Electrocardiographic Findings in Brazilian Adults without Heart Disease: ELSA-Brasil. Arq
Bras Cardiol. 2017 Nov;109(5):416-24. PubMed PMID: 28977056. Pubmed Central
PMCID: 5729777.
36. Cardiologia SBd. www.cardiometro.com.br 2019 [cited 2019 09/08/2019].
37. Evaluation IfHMa. http://www.healthdata.org/data-visualization/gbd-compare 2019
[cited 2019 09/08/2019].
38. Association AHAaAS. CARDIOVASCULAR DISEASE: A COSTLY BURDEN
FOR AMERICA PROJECTIONS THROUGH 2035 (AHA CVD Burden Report). 2017.
39. Stevens B, Pezzullo L, Verdian L, Tomlinson J, George A, Bacal F. The Economic
Burden of Heart Conditions in Brazil. Arq Bras Cardiol. 2018 Jul;111(1):29-36. PubMed
PMID: 30110042. Pubmed Central PMCID: 6078379.
40. Association AH. How to Help Prevent Heart Disease at Any Age
(heart.org/en/healthy-living/healthy-lifestyle/how-to-help-prevent-heart-disease-at-any-
age) 2019 [cited 2019 12/08/2019].
41. Khan SS, Ning H, Shah SJ, Yancy CW, Carnethon M, Berry JD, et al. 10-Year Risk
Equations for Incident Heart Failure in the General Population. Journal of the American
College of Cardiology. 2019 May 21;73(19):2388-97. PubMed PMID: 31097157. Pubmed
Central PMCID: 6527121.
42. Deen JF, Rhoades DA, Noonan C, Best LG, Okin PM, Devereux RB, et al.
Comparison of QRS Duration and Associated Cardiovascular Events in American Indian
Men Versus Women (The Strong Heart Study). The American journal of cardiology. 2017
Jun 1;119(11):1757-62. PubMed PMID: 28416200.
43. Shah AJ, Vaccarino V, Janssens AC, Flanders WD, Kundu S, Veledar E, et al. An
Electrocardiogram-Based Risk Equation for Incident Cardiovascular Disease From the
National Health and Nutrition Examination Survey. JAMA cardiology. 2016 Oct 1;1(7):779-
86. PubMed PMID: 27487404.
44. Groot A, Bots ML, Rutten FH, den Ruijter HM, Numans ME, Vaartjes I.
Measurement of ECG abnormalities and cardiovascular risk classification: a cohort study of
primary care patients in the Netherlands. The British journal of general practice : the journal
of the Royal College of General Practitioners. 2015 Jan;65(630):e1-8. PubMed PMID:
25548311. Pubmed Central PMCID: 4276002.
45. Kligfield P, Gettes LS, Bailey JJ, Childers R, Deal BJ, Hancock EW, et al.
Recommendations for the standardization and interpretation of the electrocardiogram: part
I: The electrocardiogram and its technology: a scientific statement from the American Heart
Association Electrocardiography and Arrhythmias Committee, Council on Clinical
Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm
Society: endorsed by the International Society for Computerized Electrocardiology.
Circulation. 2007 Mar 13;115(10):1306-24. PubMed PMID: 17322457.
46. Macfarlane P. Comprehensive Electrocardiology. 2 ed. Springer, editor2011.
47. Blackburn H KA, Simonson E, Rautaharju P, Punsar S. The electrocardiogram in
population studies a classification system. Circulation. 1960;21:1160-75.
48. Keys A WP. Cardiovascular Epidemiology. New York, NY: Paul B Hoeber
1956;1:22.
49. Keys A TH, Blackburn H, Brozek J, Anderson JT, Simonson E. . Coronary heart
disease among Minnesota business and professional men followed 15 years. Circulation.
1963;28:381-95.
88
80. Luepker RV, Apple FS, Christenson RH, Crow RS, Fortmann SP, Goff D, et al. Case
definitions for acute coronary heart disease in epidemiology and clinical research studies: a
statement from the AHA Council on Epidemiology and Prevention; AHA Statistics
Committee; World Heart Federation Council on Epidemiology and Prevention; the European
Society of Cardiology Working Group on Epidemiology and Prevention; Centers for Disease
Control and Prevention; and the National Heart, Lung, and Blood Institute. Circulation. 2003
Nov 18;108(20):2543-9. PubMed PMID: 14610011.
91
ANEXOS, APÊNDICE
ANEXO A
ANEXO B
Participação no estudo:
O(a) Sr./Sra. é convidado(a) a participar da pesquisa Elsa, a qual envolve o acompanhamento
dos participantes por pelo menos sete anos, com a realização de entrevistas, de exames e
medidas que ocorrerão no Centro de Investigação Elsa-MG, situado no Campus Saúde da
UFMG, Av. Alfredo Balena, 190.
Na primeira visita, sua permanência no Centro deverá ser de cerca de quatro horas, e caso
necessário, será solicitada sua liberação para participar da pesquisa em horário de trabalho.
Nessa oportunidade, o(a) Sr./Sra. será entrevistado sobre questões do seu dia a dia e
1
Universidade Federal de Minas Gerais (UFMG), Universidade de São Paulo (USP), Fundação Osvaldo
Cruz (Fiocruz) e Universidade Estadual do Rio de Janeiro (UERJ), Universidade Federal da Bahia (UFBA).
Universidade Federal do Espírito Santo (UFES) e Universidade Federal do Rio Grande do Sul (UFRGS).
93
condições de saúde, realizará algumas medidas (peso, altura e pressão arterial), exame de
urina, ultrassom do abdome e do coração, eletrocardiograma, e fotografia do fundo de olho.
Realizará também exames de sangue2, para os quais serão feitas duas coletas: a primeira
quando chegar, em jejum de 12 horas, e a segunda, após duas horas da ingestão de uma
bebida doce padrão. O total de sangue coletado será de 80 ml (cerca de 5 colheres de sopa),
uma quantidade muito menor do que uma doação de sangue, e que para uma pessoa adulta,
não traz inconveniências. O procedimento para a coleta de sangue segue rotinas hospitalares
e espera-se apenas um leve desconforto associado à picada da agulha. Algumas vezes pode
haver sensação momentânea de desmaio ou pequena reação local, mas esses efeitos são
passageiros e não oferecem riscos. A maioria dos exames que será realizada faz parte da
rotina médica e nenhum deles emite radiação ou representa risco para gestantes.
Após esta primeira visita, faremos contatos breves durante os próximos anos para
acompanhar as modificações no seu estado de saúde. O(a) Sr./Sra. será contatado por
telefone, correspondência ou e-mail, para obtenção de informações adicionais e também será
convidado para novas visitas ao Centro de Investigação Elsa-MG, que ocorrerão a cada três
anos. Por isso, é muito importante informar seu novo endereço e telefone em caso de
mudança.
Para poder acompanhar melhor sua situação de saúde, solicitaremos informações à UFMG
e a outras instituições do sistema de saúde, sobre a ocorrência de hospitalizações, licenças
2
Hemograma, exames diagnósticos para diabetes (p. ex. glicose em jejum e teste de tolerância à glicose),
creatinina, dosagem de lipídios, hormônios associados ao diabetes ou à doença cardiovascular e provas de
atividade inflamatória.
94
Não será feito nenhum pagamento pela sua participação, porém os resultados das medidas e
exames realizados que sejam relevantes para sua avaliação clínica lhe serão entregues. O
Sr./Sra. também poderá obter esclarecimentos sobre todas as informações que foram
coletadas, mediante solicitação, a qualquer momento. Caso sejam detectadas alterações, se
necessário, o(a) Sr./Sra. será encaminhado(a) com orientação para acompanhamento em
unidades da rede SUS ou outro serviço de saúde de sua preferência. O Centro de Investigação
Elsa-MG estará à disposição para prestar esclarecimentos sobre os exames e seus resultados
aos profissionais médicos envolvidos. Se durante a sua permanência no Centro de
95
Todas as informações obtidas das entrevistas e nos exames serão confidenciais, identificadas
por um número e sem menção ao seu nome. Elas serão utilizadas exclusivamente para fins
de análise estatística, e serão guardadas com segurança - somente terão acesso a elas os
pesquisadores envolvidos no projeto, ou profissionais de saúde que porventura venham a lhe
prestar atendimento ou cuidados como participante. Isso significa que a informação coletada
é sigilosa, mas não anônima, podendo ser identificada quando necessário, pelo nome do
participante, a partir de um código identificador do projeto. Contudo, em nenhuma hipótese
será permitido aos seus empregadores ou superiores hierárquicos o acesso às informações
coletadas.
Uma cópia deste Termo de Consentimento lhe será entregue. Se houver perguntas ou
necessidade de mais informações sobre o estudo ou qualquer intercorrência o Sr poderá
contatar a Professora Dra. Sandhi Maria Barreto, Av. Alfredo Balena 190, sala 813, telefone
(31) 32489938, coordenadora do projeto no Centro de Investigação Elsa-MG. O Comitê de
Ética e Pesquisa da Universidade Federal de Minas Gerais pode ser contatado pelo seguinte
telefone: (31) 2409 4592 . Sua assinatura abaixo significa que o(a) Sr./Sra. leu e
compreendeu todas as informações e concorda em participar da pesquisa ELSA.
96
Assinatura___________________________________________
Declaro concordar que amostras de sangue sejam armazenadas para análises futuras sobre as
doenças em estudo.
Sim Não
Sim Não
Assinatura______________________________________
Local_____________________________Data _______/_______/_______
Assinatura: _______________________________
Testemunha: _________________________