body-mass index in 2.3 million adolescents and cardiovascular death in adulthood

8/18/2019 Body-Mass Index in 2.3 Million Adolescents and Cardiovascular Death in Adulthood

1/11

T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e

n engl j med nejm.org 1

From the Department of Medicine (G.T.,A.L., D.B.-A.S., A.S.) and the Dr. PinchasBornstein Talpiot Medical LeadershipProgram (G.T., G.Y.), Sheba Medical Cen-ter, Tel Hashomer, the Israel DefenseForces Medical Corps (G.T., G.Y., A.L.,E.D., D.T.), Sackler School of Medicine,Tel Aviv University, Tel Aviv (G.T., A.L.,E.D., D.B.-A.S., A.A., A.S.), and HebrewUniversity–Hadassah School of PublicHealth and Community Medicine (H.L.,

J.D.K.) and the Israel Ministry of Health(N.G., A.A., Z.H.), Jerusalem — all in Is-rael; and the Department of Medicine,Mount Auburn Hospital, Harvard Medi-cal School, Cambridge, MA (A.L.). Ad-dress reprint requests to Dr. Twig at theDepartment of Medicine B, Sheba Medi-cal Center, Tel Hashomer, Ramat Gan52621, Israel, or at gilad [email protected].

This article was published on April 13,2016, at NEJM.org.

DOI: 10.1056/NEJMoa1503840Copyright © 2016 Massachusetts Medical Society.

BACKGROUND

In light of the worldwide increase in childhood obesity, we examined the associa-tion between body-mass index (BMI) in late adolescence and death from cardio-

vascular causes in adulthood.

METHODS

We grouped data on BMI, as measured from 1967 through 2010 in 2.3 million Israeliadolescents (mean age, 17.3±0.4 years), according to age- and sex-specific percen-tiles from the U.S. Centers for Disease Control and Prevention. Primary outcomes

were the number of deaths attributed to coronary heart disease, stroke, sudden deathfrom an unknown cause, or a combination of all three categories (total cardiovas-cular causes) by mid-2011. Cox proportional-hazards models were used.

RESULTS

During 42,297,007 person-years of follow-up, 2918 of 32,127 deaths (9.1%) were

from cardiovascular causes, including 1497 from coronary heart disease, 528 fromstroke, and 893 from sudden death. On multivariable analysis, there was a gradedincrease in the risk of death from cardiovascular causes and all causes that start-ed among participants in the group that was in the 50th to 74th percentiles of BMI(i.e., within the accepted normal range). Hazard ratios in the obese group (≥95thpercentile for BMI), as compared with the reference group in the 5th to 24th per-centiles, were 4.9 (95% confidence interval [CI], 3.9 to 6.1) for death from coronaryheart disease, 2.6 (95% CI, 1.7 to 4.1) for death from stroke, 2.1 (95% CI, 1.5 to2.9) for sudden death, and 3.5 (95% CI, 2.9 to 4.1) for death from total cardiovas-cular causes, after adjustment for sex, age, birth year, sociodemographic charac-teristics, and height. Hazard ratios for death from cardiovascular causes in thesame percentile groups increased from 2.0 (95% CI, 1.1 to 3.9) during follow-upfor 0 to 10 years to 4.1 (95% CI, 3.1 to 5.4) during follow-up for 30 to 40 years;during both periods, hazard ratios were consistently high for death from coronaryheart disease. Findings persisted in extensive sensitivity analyses.

CONCLUSIONS

A BMI in the 50th to 74th percentiles, within the accepted normal range, duringadolescence was associated with increased cardiovascular and all-cause mortalityduring 40 years of follow-up. Overweight and obesity were strongly associated withincreased cardiovascular mortality in adulthood. (Funded by the Environment andHealth Fund.)

A B S T R A C T

Body-Mass Index in 2.3 Million Adolescentsand Cardiovascular Death in Adulthood

Gilad Twig, M.D., Ph.D., Gal Yaniv, M.D., Ph.D., Hagai Levine, M.D., M.P.H.,Adi Leiba, M.D., M.H.A., Nehama Goldberger, M.Sc., Estela Derazne, M.Sc.,

Dana Ben-Ami Shor, M.D., Dorit Tzur, M.B.A., Arnon Afek, M.D., M.H.A.,Ari Shamiss, M.D., M.P.H., Ziona Haklai, M.A., and Jeremy D. Kark, M.D., Ph.D.

Original Article

The New England Journal of MedicineDownloaded from nejm.org by R LOPEZ MATA on April 13, 2016. For personal use only. No other uses without permission.

Copyright © 2016 Massachusetts Medical Society. All rights reserved.


2/11



Overweight and obesity in adoles-cents have increased substantially in re-cent decades and affect a third of theadolescent population in some developed coun-tries.1 Obesity early in life is considered to be arisk factor for death from cardiovascular diseaseand from all causes in adulthood, 2-5 although notall investigators concur6; such obesity may limitthe increase in life expectancy that otherwise

would be achieved.7,8 Despite progress in preven-tion and treatment of cardiovascular disease,cardiovascular mortality among young adultseither has not declined or the decline has slowedover recent decades in several developed coun-tries coincident with the obesity epidemic.9,10 Some,2,3,11 although not all, 12 studies suggest thata body-mass index (BMI, the weight in kilogramsdivided by the square of the height in meters)that falls within the upper-normal range in ado-lescence is associated with an increased risk ofdeath from cardiovascular causes, although a de-termination of the BMI threshold that is associ-ated with such an increased risk of remains un-certain. Thus, our main objective was to assessthe risk of fatal cardiovascular events in adult-hood according to the BMI range during adoles-cence. We also estimated the predicted propor-tion of cardiovascular deaths that could beattributed to the increasing prevalence of over-

weight and obesity among adolescents.

Our study was based on a national databaseof 2.3 million Israeli adolescents in whom heightand weight were measured between 1967 and2010. We assessed the association between theBMI in late adolescence and death from coro-nary heart disease, stroke, and sudden death inadulthood.

M e t h o d s

Study Population

One year before mandatory military service, Is-

raeli adolescents who are 17 years of age arerequired to undergo a medical evaluation (Fig. 1).Of the 2,454,693 adolescents between the ages of16 and 19 years who were examined from Janu-ary 1, 1967, to December 31, 2010, we excluded64,186 with missing weight and height measure-ments and 92,377 who were from non-Jewishpopulations, since these adolescents were notrepresentative of the overall minority population,in which residents are mostly exempt from mili-

tary service and therefore are not routinelycalled up for an obligatory health examination.These exclusions resulted in a study sample of2,298,130 participants, which included a nation-ally representative group of Jewish men but notof Jewish women, since Orthodox women arealso exempt from military service.

Mortality Outcomes and Documentation

of Cause of Death

The primary outcomes of the study were deathsthat had occurred by June 30, 2011, and that

were attributed to coronary heart disease, stroke,sudden death from an unknown cause, or a com-bination of the three categories (total cardiovas-cular causes). Codes for causes of death fromthe International Classification of Diseases (ICD) areprovided in Figure 1. We also evaluated deathsfrom noncardiovascular causes and all causes.We linked the underlying cause of death, as of-ficially coded from death notifications by theIsrael Central Bureau of Statistics according tothe ICD ninth revision (ICD-9) (1981–1997) andtenth revision (ICD-10) (1998–2011), to the mili-tary health-examination database; the officialcause of death was unavailable before 1981.Deaths among Israel Defense Forces personnelhave been recorded since 1967 with a notationas to whether the death was service-related.

Data Collection and Study Variables

We calculated the BMI from weight and heightmeasured by means of a beam balance and sta-diometer with each participant barefoot and

wearing underwear. Physicians reviewed medicalrecords, performed the health examinations, andprovided standardized diagnostic codes whereapplicable. Data regarding education, socioeco-nomic status, and country of origin were recorded.

The age at the time of examination, year ofbirth, and height were treated as continuous

variables. Education was grouped according to

whether the participant attended formal school-ing for less than 9 years, 10 years, 11 years, or12 years. Socioeconomic status was estimated ac-cording to residential locality,13 with participantsgrouped into low, medium, and high categories.The place of origin was defined as the birthplaceof the father or grandfather (if the father wasborn in Israel) and categorized according to coun-try or region. BMI values were grouped accord-ing to percentiles for age and sex established by




3/11


BMI in Adolescents and Cardiovascular Death

the U.S. Centers for Disease Control and Preven-tion (CDC)14 as follows: less than 5th percentile(underweight), 5th to 24th percentile (referencegroup), 25th to 49th percentile, 50th to 74th per-centile, 75th to 84th percentile, 85th to 94th per-centile (overweight), and 95th percentile or higher(obese) (Table 1). We also analyzed absolute BMI

values.

Study Oversight and Conduct

The study was funded by a research grant fromthe Environment and Health Fund in Jerusalem.The first and last authors designed the study,

wrote the first draft of the manuscript, incorpo-rated revisions from the coauthors, and madethe decision to submit the manuscript for publi-cation. The institutional review board of the Is-rael Defense Forces Medical Corps approved thestudy and waived the requirement for writteninformed consent.

Statistical Analysis

We calculated unadjusted rates of death perperson-year and Kaplan–Meier survival curvesfor the BMI percentile categories. We used Coxproportional-hazards models to estimate thehazard ratios and 95% confidence intervals forthe cardiovascular-specific outcomes. In theseanalyses, we compared the CDC percentile cat-egories with the reference group, which included

BMI values in the 5th through 24th percentiles.Time to event was modeled. Follow-up began atthe time of the original medical examinationand concluded on the date of death from anycause or June 30, 2011, whichever came first.

Included in multivariable models as potentialconfounders were all available variables that

were known to be associated with cardiovascu-lar outcomes and with BMI and that were con-firmed in our study (age, birth year, sex, educa-tion, socioeconomic status, and country of origin)as well as height, which was not confirmed as

an association in our study (Table S1 in theSupplementary Appendix, available with the fulltext of this article at NEJM.org). We included age,birth year, and sex in model 1 and all the vari-ables in model 2. We imputed missing values forcountry of origin, education, and socioeconomicstatus (1.4% of the cohort) using multiple impu-tation algorithms (SAS Miner, version 13.1). Theassumption of proportionality of the hazards wasconfirmed for all variables.

We applied the PHREG procedure to accountfor competing risks using the Cox proportional-hazards model (PROC PHREG, SAS software, ver-

sion 9.4). In tests for trend, we applied linear andquadratic terms using the midpoint of each BMIpercentile group or the median of each absoluteBMI group. Spline models (SmoothHR and sur-

vival packages, R software, version 3.2.2) were fitto estimate the BMI value associated with mini-mum cardiovascular mortality and the lowest BMIassociated with significantly increased mortali-ty. Details regarding extensive sensitivity analy-ses are provided in the Supplementary Appendix.

Figure 1. Study Design and Mortality Outcomes.Officially coded underlying causes of death were available starting in 1981,whereas data regarding service-related deaths were available from theIsrael Defense Forces throughout the full study period (1967 through2011). The body-mass index (BMI) among adolescents between the agesof 16 and 19 years was determined from measurements of weight andheight during medical examinations as part of their evaluation for servicein the Israeli military. Codes in the International Classification of Diseases (ICD) referring to the primary cardiovascular outcomes are indicated inparentheses. The underlying cause of death was coded from death notifi-cations by the Israel Central Bureau of Statistics, according to the ICD 9threvision (ICD-9) from 1981 through 1997 and 10th revision (ICD-10) from1998 through 2011.

2,298,130 Were included in the analysis

2,454,693 Were assessed from1967 through 2010

156,563 Were excluded64,186 Had missing

BMI data92,377 Were non-Jewish

minorities who wereunrepresentative of the minority population

3,991 Died between 1967 and 1980803 Had missing cause of death

32,127 Died between 1981 and 20111497 Had coronary heart disease

(ICD-9 codes, 410–414; ICD-10codes, 120–125)528 Had a stroke (ICD-9 codes,

430–4 34 and 436–438; ICD-10codes, 160–169)

893 Had sudden death (ICD-9 code,798; ICD-10 code, R96)

1662 Had missing cause of death

36,118 Died between 1967 and 2011




4/11



T a b

l e 1 . C h a r a c

t e r i s t

i c s o

f t h e P a r t

i c i p a n

t s a t

B a s e l

i n e ,

A c c o r

d i n g

t o P e r c e n

t i l e o

f B o

d y -

M a s s

I n d e x

( B M I )

. *

V a r

i a b l e

B M I P e r c e n

t i l e

< 5 t h

( N = 1 5 0 , 4 8 8 )

5 t h t o 2 4 t h

( N = 4 7 2 , 2 2 6 )

2 5 t h t o 4 9 t h

( N = 6 0 0 , 9 6 6 )

5 0 t h t o 7 4 t h

( N = 5 7 7 , 3 3 1 )

7 5 t h t o 8 4 t h

( N = 2 1 6 , 0

8 6 )

8 5 t h t o 9 4 t h

( N = 1 9 4 , 9 7 2 )

≥ 9 5 t h

( N = 8 6 , 0

6 1 )

A l l P e r c e n t

i l e s

( N = 2 , 2 9 8 , 1 3 0 )

A g e — y r

1 7 . 4

± 0 . 5

1 7 . 4

± 0 . 5

1 7 . 3

± 0 . 4

1 7 . 3

± 0 . 4

1 7 . 3 ± 0 . 4

1 7 . 3

± 0 . 4

1 7 . 3

± 0 . 4

1 7 . 3

± 0 . 4

M a l e s e x — n

o . ( % )

1 0 9 , 1 3 8 ( 7 3 )

3 0 2 , 5 2 7 ( 6 4 )

3 5 3 , 2 6 5 ( 5 9 )

3 2 0 , 5 1 7 ( 5 6 )

1 1 6 , 2 3 3 ( 5 4 )

1 0 8 , 8 2 0 ( 5 6 )

5 9 , 5

2 0 ( 6 9 )

1 , 3 7 0 , 0 2 0 ( 6 0 )

M e d

i a n

B M I ( I Q R )

M e n

1 7 . 2

( 1 6 . 7 – 1 7 . 6

) 1 8 . 9

( 1 8 . 5 – 1 9 . 4

) 2 0 . 6

( 2 0 . 2 – 2 1 . 0

) 2 2 . 4

( 2 1 . 9 – 2 3 . 0

) 2 4 . 3 ( 2

3 . 9 – 2 4 . 7

) 2 6 . 4

( 2 5 . 7 – 2 7 . 3

) 3 0 . 8

( 2 9 . 4 – 3 2 . 9

) 2 1 . 1

( 1 9 . 4 – 2 3 . 1

)

W o m e n

1 6 . 8

( 1 6 . 3 – 1 7 . 1

) 1 8 . 4

( 1 8 . 0 – 1 8 . 9

) 2 0 . 1

( 1 9 . 7 – 2 0 . 6

) 2 2 . 1

( 2 1 . 5 – 2 2 . 8

) 2 4 . 2 ( 2

3 . 9 – 2 4 . 8

) 2 6 . 8

( 2 6 . 0 – 2 7 . 9

) 3 2 . 0

( 3 0 . 8 – 3 4 . 1

) 2 1 . 1

( 1 9 . 4 – 2 3 . 2

)

B M I r a n g e

†

M e n

1 2 . 1

1 – 1 7 . 8

4

1 7 . 8

5 – 1 9 . 7

1

1 9 . 7

2 – 2 1 . 4

1

2 1 . 4

2 – 2 3 . 6

2

2 3 . 6

3 – 2 5 . 1

2

2 5 . 1

3 – 2 8 . 4

3

2 8 . 4

4 – 4 7 . 5

4

1 2 . 1

1 – 4 7 . 5

4

W o m e n

1 2 . 5

5 – 1 7 . 3

1

1 7 . 3

2 – 1 9 . 2

0

1 9 . 2

1 – 2 1 . 0

0

2 1 . 0

1 – 2 3 . 5

1

2 3 . 5

2 – 2 5 . 3

3

2 5 . 3

4 – 2 9 . 8

1

2 9 . 8

2 – 4 7 . 3

2

1 2 . 5

5 – 4 7 . 3

2

H e i g h

t — c

m

M e n

1 7 3 . 1 ± 7 . 1

1 7 3 . 3 ± 6 . 8

1 7 3 . 4 ± 6 . 7

1 7 3 . 6 ± 6 . 8

1 7 3 . 8

± 6 . 8

1 7 4 . 0 ± 6 . 9

1 7 4 . 1 ± 7 . 3

1 7 3 . 5 ± 6 . 9

W o m e n

1 6 3 . 2 ± 6 . 5

1 6 2 . 5 ± 6 . 1

1 6 2 . 1 ± 6 . 0

1 6 1 . 8 ± 6 . 0

1 6 1 . 8

± 6 . 1

1 6 1 . 8 ± 6 . 2

1 6 2 . 3 ± 6 . 4

1 6 2 . 1 ± 6 . 1

C o m p l e t e d h i g h s c h o o l — %

7 5

7 8

8 0

8 2

8 3

8 2

8 1

8 0

L o w s o c i o e c o n o m

i c s t a t u s — %

2 6

2 5

2 4

2 4

2 4

2 5

2 7

2 4

C o u n

t r y o r r e g i o n o f o r

i g i n — %

I s r a e l

5

6

6

6

6

6

7

6

F o r m e r

U S S R

1 2

1 2

1 2

1 3

1 4

1 4

1 6

1 3

A s i a

3 2

2 8

2 5

2 3

2 2

2 2

2 1

2 5

N o r

t h A f r i c a

2 2

2 4

2 5

2 5

2 5

2 5

2 6

2 4

E u r o p e

2 5

2 8

3 1

3 2

3 3

3 2

3 0

3 1

E t h i o p

i a

4

2

1

< 1

< 1

< 1

< 1

1

* P l u s – m

i n u s v a

l u e s a r e m e a n s ±

S D . T

h e B M I i s

t h e w e i g h

t i n

k i l o g r a m s d i v

i d e d

b y t h e s q u a r e o f

t h e

h e i g h t i n m e t e r s . I

Q R d e n o

t e s

i n t e r q u a r t

i l e r a n g e , a n

d U S S R U n i o n o f

S o v i e t

S o c i a l i s t R e p u b l i c s .

† L i s t e d a r e t h e B M I r a n g e s o f t h e

p a r t i c i p a n t s a t a m e a n a g e o f 1 7 . 3

y e a r s . F o r e a c h o f t h e 4 8 m o n t h s b e t w e e n t h e a g e s o f 1 6 . 0

y e a r s a n d 1 9 . 9

9 y e a r s , t h e r e i s a d i f f e r e n t a g e - a n d s e x -

s p e c

i f i c

B M I r a n g e

t h a t

d e t e r m i n e s

t h e p e r c e n

t i l e o f

B M I , a s m e a s u r e

d b y t h e

U . S . C

e n t e r s

f o r D

i s e a s e

C o n t r o

l a n d P r e v e n t

i o n

( a t o t a l o f 9 6 i n t h i s s t u d y ) .




5/11



Of the 36,118 deaths that were recorded be-tween 1967 and 2011, a total of 3991 occurredbetween 1967 and 1980. Of these deaths, 3188

were attributed to causes related to military service, and 803 had a missing underlying cause.The cause of death was also missing for 1662 ofthe 32,127 deaths (5.2%) that occurred from1981 through 2011, predominantly among youngadults (of whom 48% were


6/11


7/11



S4 in the Supplementary Appendix). Participantsin the BMI group categorized as overweight (85thto 94th percentiles) had hazard ratios of 3.0 (95%CI, 2.5 to 3.7) for death from coronary heart dis-ease, 1.8 (95% CI, 1.3 to 2.5) for death fromstroke, 1.5 (95% CI, 1.1 to 1.9) for sudden death,and 2.2 (95% CI, 1.9 to 2.6) for death from totalcardiovascular causes. The hazard ratios weregenerally lowest for all categories of death amongparticipants in the BMI groups that were in the25th to 49th or the 5th to 24th BMI percentiles.

In calculations of the risk of death from totalcardiovascular causes at different follow-up timesand 10-year intervals, the association betweencardiovascular mortality and increased BMI wasevident by 10 years of follow-up (hazard ratio,2.0; 95% CI, 1.1 to 3.9) and was more pro-nounced during the follow-up period from 30 to40 years (hazard ratio, 4.1; 95% CI, 3.1 to 5.4)(Table S5 and Fig. S3 in the Supplementary Ap-pendix). The risk was persistently high for deathfrom coronary heart disease across the entireperiod of follow-up.

In the analysis of absolute values of BMI,participants with a BMI ranging from 20.0 to22.4 had a higher risk of death from coronaryheart disease than did those with a BMI rangingfrom 17.5 to 19.9 (hazard ratio, 1.2; 95% CI, 1.1to 1.4) (Table S6 in the Supplementary Appen-dix). The risks of death from stroke, sudden

death, and total cardiovascular causes (along with noncardiovascular causes and all causes) were also elevated starting with a BMI of 22.5and increased more steeply among the extremelyobese for cardiovascular-specific death. We usedmultivariable-adjusted spline models to estimatethat the minimum risks of death from stroke,sudden death, and cardiovascular causes wereamong participants who had BMI values of 19.8,19.3, and 18.3, respectively, whereas the associa-tion with death from coronary artery disease

was graded across the full range of BMI-percen-

tile groups. Participants had a significantly elevated risk of death from total cardiovascularcauses starting at BMI values above 20.3 (Fig. S4in the Supplementary Appendix). To compareour data with previous observations, 2 we com-puted z scores of BMI for the 1,854,522 partici-pants who were 17 or 18 years of age when they

were evaluated. After multivariable adjustment,the hazard ratios for death from coronary heartdisease were 1.54 (95% CI, 1.46 to 1.62) for men

and 1.58 (95% CI, 1.31 to 1.91) for women per1-unit increment in the z score.

The findings persisted in a series of sensitiv-ity analyses, including a sex-specific analysis(which showed overall consistency between thesexes), an analysis restricted to adolescents withunimpaired health status, an analysis modelingthe association without competing risks thatused the group with BMI values in the 25th to49th percentiles as the reference category, analy-ses that were restricted to participants who wereevaluated after 1981 and to those who wereevaluated from 1967 through 1980, an analysiscomparing deaths that occurred before the ageof 45 years with those that occurred at the ageof 45 years or older, analyses according to calen-dar periods of follow-up and country of origin,and an analysis that included all cardiovasculardiseases defined according to ICD coding (9.9%of all deaths) (Table S4 and Tables S7 throughS13 in the Supplementary Appendix). Simulationof deaths with an unknown underlying cause

with varying assignment as cardiovasculardeaths was shown to have a minor effect on theassociation between BMI and the study out-comes (Tables S14 and S15 in the SupplementaryAppendix). In an analysis of population-attribut-able fractions, among participants with a BMIhigher than the 50th percentile in 2013, theprojected population-attributable fractions were

28% for death from total cardiovascular causesand 36% for death from coronary heart disease(Table S2 in the Supplementary Appendix).

D i s c u s s i o n

In this nationwide population-based study, wefound an association between the BMI in lateadolescence and subsequent cardiovascular mor-tality, predominantly in midlife, since the cohortdid not include participants who had reached theolder ages at which cardiovascular disease is a

dominant cause of death. BMI, including mea-surements in the currently accepted mid-normalrange, was associated with a graded increase inthe risk of death from cardiovascular causes.Obesity during adolescence was associated witha substantially increased risk of cardiovascularoutcomes in middle age, particularly death fromcoronary heart disease. The associations, which

were similarly evident in both sexes, persistedstrongly for cardiovascular deaths occurring




8/11



T a b

l e 3 . H a z a r

d R a t

i o s

f o r

C a u s e

o f D e a

t h , A

c c o r d

i n g

t o P e r c e n

t i l e o

f B M I d u r i n g

A d o l e s c e n c e . *

V a r

i a b l e

B M I P e r c e n

t i l e d u r i n g

A d o l e s c e n c e

< 5 t h

5 t h t o 2 4 t h

2 5 t h t o 4 9 t h

5 0 t h t o 7 4 t h

7 5 t h t o 8 4 t h

8 5 t h t o 9 4 t h

≥ 9 5 t h

C a u s e o f

d e a

t h

C o r o n a r y h e a r t d i s e a s e

H a z a r

d r a

t i o ( 9 5 % C I ) † 0 . 9

5 ( 0 . 7

3 – 1 . 2 4 )

1 . 0 0 ( r e f e r e n c e

)

1 . 1 1 ( 0 . 9

4 – 1 . 3 1 )

1 . 4 9 ( 1 . 2

7 – 1 . 7 6 )

2 . 1 7 ( 1 . 7

8 – 2 . 6 4 )

3 . 0 2 ( 2 . 5

0 – 3 . 6 5 )

4 . 8 9 ( 3 . 9

1 – 6 . 1 2 )

P v a

l u e

0 . 7 2

0 . 2 3

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

S t r o

k e H a z a r

d r a

t i o ( 9 5 % C I ) † 0 . 9

4 ( 0 . 6

3 – 1 . 4 1 )

1 . 0 0 ( r e f e r e n c e

)

0 . 8 4 ( 0 . 6

5 – 1 . 1 0 )

1 . 1 8 ( 0 . 9

2 – 1 . 5 3 )

1 . 4 2 ( 1 . 0

3 – 1 . 9 7 )

1 . 8 1 ( 1 . 3

0 – 2 . 5 1 )

2 . 6 4 ( 1 . 7

2 – 4 . 0 8 )

P v a

l u e

0 . 7 7

0 . 2 1

0 . 1 9

0 . 0 3 4

< 0 . 0

0 1

< 0 . 0

0 1

S u d d e n d e a t h

H a z a r

d r a

t i o ( 9 5 % C I ) † 1 . 1

5 ( 0 . 8

8 – 1 . 5 0 )

1 . 0 0 ( r e f e r e n c e

)

0 . 9 9 ( 0 . 8

1 – 1 . 2 0 )

1 . 1 7 ( 0 . 9

6 – 1 . 4 2 )

1 . 4 1 ( 1 . 0

9 – 1 . 8 2 )

1 . 4 6 ( 1 . 1

1 – 1 . 9 1 )

2 . 0 9 ( 1 . 5

0 – 2 . 9 1 )

P v a

l u e

0 . 3 3

0 . 9 1

0 . 1 2

0 . 0 0 9

0 . 0 0 6

< 0 . 0

0 1

T o t a l c a r

d i o v a s c u

l a r c a u s e s

H a z a r

d r a

t i o ( 9 5 % C I ) † 1 . 0

2 ( 0 . 8

6 – 1 . 2 1 )

1 . 0 0 ( r e f e r e n c e

)

1 . 0 1 ( 0 . 9

0 – 1 . 1 4 )

1 . 3 2 ( 1 . 1

8 – 1 . 4 8 )

1 . 7 6 ( 1 . 5

3 – 2 . 0 3 )

2 . 2 5 ( 1 . 9

6 – 2 . 5 8 )

3 . 4 6 ( 2 . 9

3 – 4 . 1 0 )

P v a l u e

0 . 7 9

0 . 8 1

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

N o n c a r d

i o v a s c u l a r c a u s e s

H a z a r

d r a

t i o ( 9 5 % C I )

1 . 0

5 ( 1 . 0

0 – 1 . 1 1 )

1 . 0 0 ( r e f e r e n c e

)

0 . 9 9 ( 0 . 9

5 – 1 . 0 2 )

1 . 0 4 ( 1 . 0

1 – 1 . 0 8 )

1 . 1 6 ( 1 . 1

1 – 1 . 2 1 )

1 . 2 3 ( 1 . 1

7 – 1 . 2 9 )

1 . 5 4 ( 1 . 4

4 – 1 . 6 4 )

P v a

l u e

0 . 0 4

0 . 3 7

0 . 0 1 9

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

A l l c a u s e s

H a z a r d r a t i o ( 9 5 % C I )

1 . 0

5 ( 1 . 0

0 – 1 . 1 1 )

1 . 0 0 ( r e f e r e n c e )

0 . 9 9 ( 0 . 9

6 – 1 . 0 2 )

1 . 0 6 ( 1 . 0

3 – 1 . 1 0 )

1 . 2 0 ( 1 . 1

5 – 1 . 2 6 )

1 . 3 0 ( 1 . 2

4 – 1 . 3 6 )

1 . 6 8 ( 1 . 5

8 – 1 . 7 8 )

P v a

l u e

0 . 0 3 9

0 . 4 5

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

< 0 . 0

0 1

* A t o t a l o f 2 9 1 8 p a r t i c i p a n t s d i e d f r o m c a r d i o v a s c u l a r c a u s e s ( i n c l u d i n g 1 4 9 7 f r o m c o r o n a r y h e a r t d i s e a s e , 5

2 8 f r o m s t r o k e , a

n d 8 9 3 f r o m s u d d e n d e a t h o f u n k n o w n c a u s e ) , 2

9 , 2 0 9 d i e d

f r o m n o n c a r

d i o v a s c u

l a r c a u s e s , a n

d 3 2 , 1

2 7 d i e d

f r o m a l

l c a u s e s .

H a z a r

d r a t

i o s w e r e c a

l c u l a t e d w

i t h t h e u s e o f

t h e C o x p r o p o r

t i o n a

l - h a z a r

d s m e t

h o d a m o n g 2 ,

2 9 8 , 1 3 0 p a r t i c i p a n t s

b e -

t w e e n t h e a g e s o f 1 6 a n d 1 9 y e a r s a t t h e t i m e o f e x a m i n a t i o n w i t h 4 2 , 2

9 7 , 0 0 7 p e r s o n - y e a r s o f f o l l o w - u p ; t h e h a z a r d r a t i o s w e r e a d j u s t e d f o r a g e , b

i r t h y e a r , s

e x , s o c i o e c o n o m i c s t a t u s ,

c o u n t r y o f o r i g i n , e

d u c a t i o n a l l e v e l , a n d h e i g h t . I

n t h e s e a n a l y s e s , w

e a c c o u n t e d f o r c o m p e t i n g r i s k ( d e a t h s f r o m o t h e r c a u s e s ) . I n t e s t s f o r t r e n d f o r a l l c o m p a r i s o n s , P < 0 . 0 0 1 f o r l i n e a r

t e r m s a n

d P ≤ 0 . 0 0 4 f o r q u a

d r a t i c

t e r m s .

† T h e

h a z a r d r a

t i o i n t h i s c a

t e g o r y

w a s e s s e n t

i a l l y u n c h a n g e

d w

h e n s o c i o e c o n o m

i c s t a t u s a n

d h e i g h t w e r e e x c l u d e d

f r o m

t h e s e v e n - c o v a r i a

t e m o d e l .




9/11



during four decades after the measurement ofBMI in adolescence. The association withstood aseries of sensitivity analyses. The population-attrib-utable fractions in our study are projected to rise

with the secular shift in the BMI distributionand indicate that overweight and obesity in ado-lescence may account for a fifth of cardiovascu-lar deaths and a quarter of deaths from coronaryheart disease by the time the participants reachmidlife. Modeled predictions for the United Statespoint to a marked increase in the incidence ofcoronary heart disease and in the rate of deathin young and middle adulthood. 15

Several studies2-4,11,16 (although not all 6) haveshown an association between obesity duringadolescence and a future risk of death from coro-nary heart disease, with conflicting evidence re-garding death from cerebrovascular causes. 5,6 Inthe study by Baker et al.,2 in which investigatorsevaluated the association between childhoodBMI and coronary events in adulthood, the BMIat the age of 7 years was associated with adultmortality from coronary heart disease (hazardratio, 1.10 per 1-unit increment in the z score inboys); the strength of the association increasedup to the age of 13 years (hazard ratio, 1.24). Inour study, in similar analyses restricted to ado-lescents who were 17 or 18 years of age, we es-timated hazard ratios for death from coronaryheart disease of 1.54 and 1.58 for boys and girls,

respectively, findings that appear to be consis-tent with a stronger association in adolescencethan earlier in childhood.

The large size of our study, which incorporatedmore than 42 million person-years of follow-up,provided adequate statistical power to assess theassociations within the currently accepted normalrange of BMI values. Excess all-cause mortality(including cardiovascular mortality) starting atthe 50th percentile of adolescent BMI valuesconfirmed the findings of an earlier study on aportion of this cohort. 17 Thus, the classification

of BMI according to the accepted normal range(i.e., the 5th to 84th percentiles and a BMI rang-ing from 18.5 to 25.0) may underestimate therisk associated with being overweight in adoles-cence. This inference is supported by our findingsthat there is a graded increase in the risk of deathstarting at the mid-normal range of adolescentBMI (50th to 74th percentiles) and that the high-normal BMI range (75th to 84th percentiles) wasassociated with hazard ratios of 2.2 for coronary

heart disease and 1.8 for total cardiovascularcauses.

Our findings appear to provide a link betweenthe secular trends in adolescent overweight andcoronary mortality during the past decades. Incontrast to the steep decline in the rate of deathfrom cardiovascular causes among older agegroups, cardiovascular mortality among youngadults has not decreased or the decline hasslowed in several developed countries.9,10

How might adolescent BMI affect adult car-diovascular mortality? In our study, we could notcontrol for important risk factors (e.g., smoking,exercise, and physical f itness)18 or for adult BMI.We have considered two pathways by whichadolescent BMI might influence cardiovascularoutcomes in adulthood. First, obesity may bedeleterious during adolescence, since it has beenassociated with unfavorable metabolic abnormali-ties (suggesting an indirect pathway mediated

Figure 2. Body-Mass Index (BMI) during Adolescence and SubsequentCardiovascular Mortality.Shown are rates of death from cardiovascular causes during up to 44 yearsof follow-up according to the percentile of BMI during adolescence, among2,298,130 participants who were assessed between the ages of 16 and 19years. Included in the 2918 deaths from cardiovascular causes were deathsfrom coronary artery disease and stroke and sudden deaths of unknowncause. Gray shading denotes 95% confidence intervals. The model was ad-justed for sex, age at examination, birth year, education, residential socio-economic status, country of origin, and height. Also shown are the num-bers of participants at risk, cumulative person-years, and the cumulativenumbers of deaths from cardiovascular causes through 40 years of follow-up. An additional 242 deaths from cardiovascular causes occurred between40 years and the study cutoff of 44 years.

1.4

1.2

1.0

0.8

0.2

0.4

0.6

0.00 10 20 30 40 C

u m u

l a t i v e

M o r

t a l i t y f r o m

C a r

d i o v a s c u

l a r

D i s e a s e

( % )

No. at RiskParticipants at riskCumulative person-yrCumulative cardio-

vascular deaths

1,712,01817,201,301

185

1,042,01830,718,320

609

540,63638,472,521

1,577

160,14541,926,636

2,676

75th –84th

50th–74th25th–49th5th–24th


10/11



through risk factors such as unfavorable plasmalipid or lipoprotein levels, increased blood pres-sure, impaired glucose metabolism, and insulinresistance15,19), cardiac remodeling,20 lengtheningof the QT interval,21 and formation of coronaryand aortic atherosclerotic plaques. 22 Furthermore,a mendelian randomization study showed a strongassociation between BMI and several cardio-metabolic risk markers in the young (even withinthe nonobese weight range), which effectivelyruled out uncontrolled confounding and reversecausation as explanations for their findings. 23 Increased coronary risk associated with adoles-cent obesity persisted in a subgroup in our co-hort,24 even after adjustment for BMI in youngadulthood, which suggests the importance ofthe timing of exposure to obesity during a life-time.12,25 However, other studies26,27 have notshown an independent association between BMIduring adolescence and adult cardiovascular dis-ease after adjustment for adult BMI. Second, asBMI tracks over time25,26 (and more stronglystarting at the age of 18 years 28), adult BMI maymediate the risk association. The associationbetween adult BMI and cardiovascular disease issupported by a mendelian randomization study. 29 Such a pathway might explain the associationbetween seemingly normal levels of BMI (50thto 74th percentiles) and death from total cardio-

vascular causes and between a BMI of more than

20.0 and death from coronary artery disease. Inthe absence of adult measures of BMI, we wereunable to address this question.

Our study has certain limitations. First, asnoted, in the absence of adult measures of BMI,

we could not assess an independent effect of ado-lescent BMI on death from cardiovascular diseasein adulthood. Second, cause-specific data wereunavailable to us for 6.8% of all deaths. How-ever, simulations indicated only minor effects ofmissing data on our risk estimates. Third, we

were unable to account for important cardiovas-

cular lifestyle risk factors that may confound the

BMI association, although adjustment for smok-ing had no effect in other studies, 3,25 and a men-delian randomization study showed effects forBMI that were independent of confounding. 23 Fourth, since our follow-up included participantsin early and mid-adulthood but not in the olderage groups in which cardiovascular death pre-dominates, fewer than 10% of our study partici-pants died from cardiovascular causes. Fifth,although the study sample was highly represen-tative of the Israeli adolescent male Jewish popu-lation, it was less representative of Israeli wom-en, and our findings need to be confirmed in aracially and ethnically diverse population. Fi-nally, since we were analyzing only death out-comes, we cannot determine whether BMI wasassociated with the incidence of cardiovasculardisease, death from cardiovascular disease, orboth. The strengths of the study include the useof standardized measurements of weight andheight rather than recalled values, 30 the largesample size, and the extended follow-up, whichprovided adequate power to examine narrowsubgroups of BMI percentiles.

In conclusion, an increased BMI in late ado-lescence, even within the currently accepted nor-mal range, was strongly associated with cardio-

vascular mortality in young adulthood or midlife.We could not determine whether an increasedBMI in adolescence is an independent risk factor,

is mediated by adult obesity, or both. The secularshift to the right in the distribution of adolescentBMI and the rising prevalence of overweight andobesity among adolescents may account for asubstantial and increasing future burden of car-diovascular disease, particularly coronary heartdisease.

Supported by a research grant from the Environment andHealth Fund in Jerusalem (RGA 1002, to Dr. Kark).

Disclosure forms provided by the authors are available withthe ful l text of this art icle at NEJM.org.

We thank Drs. Lawrence Freedman and Raanan Raz for theirstatist ical consultation, Drs. Noam Levine and Shmulik Tiosanofor their assistance in the preparation of the original figures,

and the anonymous reviewers for their thoughtful comments.

References

1. Ogden CL, Carroll MD, Kit BK, FlegalKM. Prevalence of childhood and adultobesity in the United States, 2011-2012.

JAMA 2014;311:806-14.2. Baker JL, Olsen LW, Sørensen TIA.Childhood body-mass index and the riskof coronary heart disease in adulthood.N Engl J Med 2007;357:2329-37.

3. Must A, Jacques PF, Dallal GE, BajemaCJ, Dietz WH. Long-term morbidity andmortality of overweight adolescents — afollow-up of the Harvard Growth Study of1922 to 1935. N Engl J Med 1992;327:1350-5.4. Franks PW, Hanson RL, Knowler WC,Sievers ML, Bennett PH, Looker HC.

Childhood obesity, other cardiovascularrisk factors, and premature death. N Engl

J Med 2010;362:485-93.5. Engeland A, Bjørge T, Søgaard AJ,Tverdal A. Body mass index in adolescencein relation to total mortality: 32-year fol-low-up of 227,000 Norwegian boys andgirls. Am J Epidemiol 2003;157:517-23.




11/11



6. Lawlor DA, Martin RM, Gunnell D, etal. Association of body mass index mea-sured in childhood, adolescence, and

young adulthood with risk of ischemicheart disease and stroke: findings from3 historical cohort studies. Am J ClinNutr 2006;83:767-73.7. Olshansky SJ, Passaro DJ, HershowRC, et al. A potential decline in life expec-

tancy in the United States in the 21st cen-tury. N Engl J Med 2005;352:1138-45.8. Fontaine KR, Redden DT, Wang C,Westfall AO, Allison DB. Years of life lostdue to obesity. JAMA 2003;289:187-93.9. O’Flaherty M, Allender S, Taylor R,Stevenson C, Peeters A, Capewell S. Thedecline in coronary heart disease mortal-ity is slowing in young adults (Australia1976-2006): a time trend analysis. Int JCardiol 2012;158:193-8.10. Ford ES, Capewell S. Coronary heartdisease mortality among young adults inthe U.S. from 1980 through 2002: con-cealed leveling of mortality rates. J AmColl Cardiol 2007;50:2128-32.11. Bjørge T, Engeland A, Tverdal A,Smith GD. Body mass index in adoles-cence in relation to cause-specific mortal-ity: a follow-up of 230,000 Norwegianadolescents. Am J Epidemiol 2008;168:30-7.12. van Dam RM, Willett WC, Manson JE,Hu FB. The relationship between over-

weight in adolescence and prematuredeath in women. Ann Intern Med 2006;145:91-7.13. Israel Central Bureau of Statistics.Characterization and classification of lo-cal authorities by the socio-economic lev-el of the population. Jerusalem: IsraelCentral Bureau of Statistics, 2006.14. Kuczmarski RJ, Ogden CL, Guo SS, et

al. 2000 CDC growth charts for the Unit-ed States: methods and development. Vi-tal Health Stat 11 2002;2002(246):1-190.15. Bibbins-Domingo K, Coxson P,Pletcher MJ, Lightwood J, Goldman L.Adolescent overweight and future adultcoronary heart disease. N Engl J Med2007;357:2371-9.16. DiPietro L, Mossberg HO, Stunkard

AJ. A 40-year history of overweight chil-dren in Stockholm: life-time overweight,morbidity, and mortality. Int J Obes RelatMetab Disord 1994;18:585-90.17. Twig G, Afek A, Shamiss A, et al. Ado-lescence BMI and trends in adulthoodmortality: a study of 2.16 million adoles-cents. J Clin Endocrinol Metab 2014;99:2095-103.18. Shah RV, Murthy VL, Colangelo LA, etal. Association of fitness in young adult-hood with survival and cardiovascularrisk: the Coronary Artery Risk Develop-ment in Young Adults (CARDIA) Study.

JAMA Intern Med 2016;176:87-95.19. Skinner AC, Perrin EM, Moss LA,Skelton JA. Cardiometabolic risks and se-

verity of obesity in children and youngadults. N Engl J Med 2015;373:1307-17.20. Mangner N, Scheuermann K, WinzerE, et al. Childhood obesity: impact on car-diac geometry and funct ion. JACC Cardio-

vasc Imaging 2014;7:1198-205.21. Güven A, Özgen T, Güngör O, AydınM, Baysal K. Association between the cor-rected QT interval and carotid artery inti-ma-media thickness in obese children.

J Clin Res Pediatr Endocrinol 2010;2:21-7.22. McGill HC Jr, McMahan CA, Herder-ick EE, Malcom GT, Tracy RE, Strong JP.Origin of atherosclerosis in childhoodand adolescence. Am J Clin Nutr 2000;72:Suppl:1307S-1315S.

23. Würtz P, Wang Q, Kangas AJ, et al.Metabolic signatures of adiposity in

young adults: Mendelian randomizationanalysis and effects of weight change.PLoS Med 2014;11(12):e1001765.24. Tirosh A, Shai I, Afek A, et al. Adoles-cent BMI trajectory and risk of diabetes

versus coronary disease. N Engl J Med2011;364:1315-25.

25. Jeffreys M, McCarron P, Gunnell D,McEwen J, Smith GD. Body mass index inearly and mid-adulthood, and subsequentmortality: a historical cohort study. Int JObes Relat Metab Disord 2003;27:1391-7.26. Gray L, Lee IM, Sesso HD, Batty GD.Body weight in early and mid-adulthoodin relation to subsequent coronary heartdisease mortality: 80-year follow-up inthe Harvard Alumni Study. Arch InternMed 2011;171:1768-70.27. Juonala M, Magnussen CG, BerensonGS, et al. Childhood adiposity, adult adi-posity, and cardiovascular risk factors.N Engl J Med 2011;365:1876-85.28. Kvaavik E, Tell GS, Klepp KI. Predic-tors and tracking of body mass indexfrom adolescence into adulthood: follow-up of 18 to 20 years in the Oslo YouthStudy. Arch Pediatr Adolesc Med 2003;157:1212-8.29. Hägg S, Fall T, Ploner A, et al. Adipos-ity as a cause of cardiovascular disease:a Mendelian randomization study. Int JEpidemiol 2015;44:578-86.30. Keith SW, Fontaine KR, Pajewski NM,Mehta T, Allison DB. Use of self-reportedheight and weight biases the body massindex-mortality association. Int J Obes(Lond) 2011;35:401-8.Copyright © 2016 Massachusetts Medical Society.

The New England Journal of MedicineDownloaded from nejm org by R LOPEZ MATA on April 13 2016 For personal use only No other uses without permission

body-mass index in 2.3 million adolescents and cardiovascular death in adulthood

Documents