ORIGINAL RESEARCH
P edic ion o Cumula i e Exposu e
o A he ogenic Lipids Du ing
Ea ly Adul hood
John T. Wilkins, MD, MSC,
a,b
Hongyan Ning, MD, MS,
b
No ina B. Allen, PHD,
b
Alexande Zheu lin, MD, MPH,
a
Nilay S. Shah, MD, MPH,
a,b
Ma hew J. Feins ein, MD,
a,b
Amanda M. Pe ak, MD,
a,c
Sadiya S. Khan, MD, MSC,
a,b
Ankee S. Bha , MD, MBA, SCM,
d
Ra i Shah, MD,
e
Venka esh Mu hy, MD, PHD,
Allan Snide man, MD,
g
Donald M. Lloyd-Jones, MD, SCM
a,b,c
ABSTRACT
BACKGROUND The abili y o a 1- ime measu emen o non–high-densi y lipop o ein choles e ol (non–HDL-C) o
low-densi y lipop o ein choles e ol (LDL-C) o p edic he cumula i e exposu e o hese lipids du ing ea ly adul hood
(age 18-40 yea s) and he associa ed a he oscle o ic ca dio ascula disease (ASCVD) isk a e age 40 yea s is no clea .
OBJECTIVES The objec i es o his s udy we e o e alua e whe he a 1- ime measu emen o non-HDL-C o LDL-C in a
young adul can p edic cumula i e exposu e o hese lipids du ing ea ly adul hood, and o quan i y he associa ion
be ween cumula i e exposu e o non-HDL-C o LDL-C du ing ea ly adul hood and he isk o ASCVD a e age 40 yea s.
METHODS We included CARDIA (Co ona y A e y Risk De elopmen in Young Adul s S udy) pa icipan s who we e ee
o ca dio ascula disease be o e age 40 yea s, we e no aking lipid-lowe ing medica ions, and had $3 measu emen s o
LDL-C and non–HDL-C be o e age 40 yea s. Fi s , we assessed he abili y o a 1- ime measu emen o LDL-C o non–
HDL-C ob ained be ween age 18 and 30 yea s o p edic he qua ile o cumula i e lipid exposu e om ages 18 o 40
yea s. Second, we assessed he associa ions be ween qua iles o cumula i e lipid exposu e om ages 18 o 40 yea s wi h
ASCVD e en s ( a al and non a al myoca dial in a c ion and s oke) a e age 40 yea s.
RESULTS O 4,104 CARDIA pa icipan s who had mul iple lipid measu emen s be o e and a e age 30 yea s, 3,995
pa icipan s me ou inclusion c i e ia and we e in he final analysis se . A 1- ime measu e o non–HDL-C and LDL-C
had excellen disc imina ion o p edic ing membe ship in he op o bo om qua iles o cumula i e exposu e (AUC:
0.93 o he 4 models). The absolu e alues o non–HDL-C and LDL-C ha p edic ed membe ship in he op qua iles wi h
he highes simul aneous sensi i i y and specifici y (highes Youden’s Index) we e >135 mg/dL o non–HDL-C
and >118 mg/dL o LDL-C; he alues ha p edic ed membe ship in he bo om qua iles we e <107 mg/dL o
non–HDL-C and <96 mg/dL o LDL-C. Indi iduals in he op qua ile o non–HDL-C and LDL-C exposu e had
demog aphic-adjus ed HRs o 4.6 (95% CI: 2.84-7.29) and 4.0 (95% CI: 2.50-6.33) o ASCVD e en s a e age 40 yea s,
espec i ely, when compa ed wi h each bo om qua ile.
CONCLUSIONS Single measu es o non–HDL-C and LDL-C ob ained be ween ages 18 and 30 yea s a e highly p edic i e
o cumula i e exposu e be o e age 40 yea s, which in u n s ongly p edic s la e -li e ASCVD e en s.
(JACC. 2024;84:961–973) © 2024 Published by Else ie on behal o he Ame ican College o Ca diology Founda ion
ISSN 0735-1097/$36.00 h ps://doi.o g/10.1016/j.jacc.2024.05.070
F om he
a
Depa men o Medicine (Ca diology), No hwes e n Uni e si y Feinbe g School o Medicine, Chicago, Illinois, USA;
b
Depa men o P e en i e Medicine (Epidemiology), No hwes e n Uni e si y Feinbe g School o Medicine, Chicago, Illinois,
USA;
c
Depa men o Pedia ics (Ca diology), No hwes e n Uni e si y Feinbe g School o Medicine, Chicago, Illinois, USA;
d
Kaise
Pe manen e, Di ision o Resea ch, Oakland, Cali o nia, USA;
e
Depa men o Medicine (Ca diology), Vande bil Uni e si y School
Lis en o his manusc ip ’s
audio summa y by
Edi o Eme i us
D Valen in Fus e on
www.jacc.o g/jou nal/jacc.
JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL.84,NO.11,2024
ª2024 PUBLISHED BY ELSEVIER ON BEHALF OF THE
AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION
The cumula i e exposu e o a he o-
genic lipids (non–high-densi y lipo-
p o ein choles e ol [non–HDL-C]
and low-densi y lipop o ein choles e ol
[LDL-C]) in young adul hood (ages 18-40
yea s) is a p ima y de e minan o mid- and
la e-li e a he oscle o ic ca dio ascula
disease (ASCVD).
1-3
The e o e, iden i ying
young adul s likely o expe ience a high cu-
mula i e exposu e o a he ogenic lipids
may enhance unde s anding o long- e m
ASCVD iskandmayin o mheal hbeha io
ecommenda ions and mo i a e heal h beha io op i-
miza ion in some o educe a he ogenic lipid expo-
su e and long- e m ASCVD isk.
Mos adul clinical p ac ice guidelines ecommend
a leas 1 measu emen o a he ogenic lipids be ween
ages 18 o 30 yea s, al hough he e is a i al need o
enhance awa eness o he impo ance o lipid mea-
su emen because only 50% o eligible young adul s
unde go his sc eening.
4,5
Al hough lipid alues ypically ise g adually
h ough adul hood o mos people, he e is consid-
e able a iabili y in bo h he pa e ns and a es o
change among indi iduals, and he abili y o a single
measu emen o a he ogenic lipids o p edic he cu-
mula i e lipid exposu e om young adul hood o
middle age has no been sys ema ically e alua ed.
6
Fu he , he absolu e alues ha define clinically
“high”o “low”lipid alues in un ea ed indi iduals
a e ypically de i ed om no ma i e dis ibu ions
om middle-aged adul pa icipan s in epidemiologic
s udies. Because he absolu e alues o a he ogenic
lipids a e, on a e age, lowe in young adul s when
compa ed wi h middle-aged adul s and hey end o
inc ease h ough adul hood, hese clinical h esholds
may no adequa ely eflec clinically ele an
a he ogenic lipid exposu e o young adul s.
7
To da e,
absolu e alues o a he ogenic lipids in young adul s
ha iden i y who is likely o ha e a high o low cu-
mula i e lipid exposu e h ough age 40 yea s ha e
no been defined. Likewise, i is no clea how well
hese absolu e alues p edic cumula i e exposu e
du ing ea ly adul li e. Consequen ly, cu en
guidelines a e ague in hei ecommenda ions o
wha lipids alues should be conside ed “high”o
“low”among young adul s.
To in o m his gap in he cu en li e a u e, we
calcula ed indi idual-le el cumula i e exposu e o
a he ogenic lipids (non–HDL-C and LDL-C) om ages
18 o 40 yea s among CARDIA (Co ona y A e y Risk
De elopmen in Young Adul s) s udy pa icipan s. We
hen assessed he abili y o 1- ime lipid measu e-
men sbe o eage30yea s op edic hecumula i e
exposu e o a he ogenic lipids du ing ea ly adul hood
( h ough age 40 yea s). We epo absolu e alues o
a he ogenic lipids wi h he highes simul aneous
sensi i i y and specifici y o p edic cumula i e
exposu e as well as he ela i e ASCVD e en isk
obse ed a e age 40 yea s o each qua ile o cu-
mula i e exposu e in ea ly adul li e. Ou findings
may in o m clinical ecommenda ions o lipid es ing
and p o ide p ac ical absolu e lipid alues o guide
discussion o isk ac o op imiza ion o young
adul s.
METHODS
THE CARDIA STUDY. In 1985 and 1986, he CARDIA
s udy ec ui ed 5,115 Black and Whi e men and
women aged 18 h ough 30 yea s om 4 si es ac oss
he Uni ed S a es: Chicago, Illinois; Bi mingham,
Alabama; Minneapolis, Minneso a; and Oakland,
Cali o nia. A incep ion, he coho was balanced by
ace (52% Black, 48% Whi e), sex (55% emale, 45%
male), educa ion (40% wi h #12 yea s, 60% wi h
>12 yea s o educa ion), and age (45% 18-24 yea s,
55% 25-30 yea s) a each cen e .
8
Pa icipan s unde wen in-pe son examina ions a
baseline (Y0) and a ollow-up yea s 2, 5, 7, 10, 15, 25,
30, and 35. Ongoing con ac is main ained wi h pa -
icipan s ia elephone, mail, o elec onically e e y
6 mon hs, wi h annual medical his o y and e en
asce ainmen be ween in-pe son examina ions.
Re en ion a es among su i ing pa icipan s a each
in-pe son ollow-up examina ion ha e been high, a
91%, 86%, 81%, 79%, 74%, 72%, 72%, 71%, and 67%
(du ing he COVID-19 pandemic), espec i ely. O e
he las 5 yea s, >90% o he su i ing coho mem-
be s ha e been di ec ly con ac ed. Follow-up o i al
SEE PAGE 974
ABBREVIATIONS
AND ACRONYMS
ASCVD =a he oscle o ic
ca dio ascula disease
BMI =body mass index
CVD =ca dio ascula disease
LDL-C =low-densi y
lipop o ein choles e ol
MI =myoca dial in a c ion
non–HDL-C =non–high-
densi y lipop o ein choles e ol
o Medicine, Nash ille, Tennessee, USA;
Depa men o Medicine (Ca diology), Uni e si y o Michigan Medical School, Ann A bo ,
Michigan, USA; and he
g
Depa men o Medicine (Ca diology), McGill Uni e si y School o Medicine, Mon eal, Quebec, Canada.
Re iew and accep ance occu ed unde D Valen in Fus e ’s e m as Edi o -in-Chie .
The au ho s a es hey a e in compliance wi h human s udies commi ees and animal wel a e egula ions o he au ho s’
ins i u ions and Food and D ug Adminis a ion guidelines, including pa ien consen whe e app op ia e. Fo mo e in o ma ion,
isi he Au ho Cen e .
Manusc ip ecei ed Decembe 4, 2023; e ised manusc ip ecei ed May 2, 2024, accep ed May 16, 2024.
Wilkins e al JACC VOL. 84, NO. 11, 2024
P edic ion o Lipid Exposu e in Young Adul s SEPTEMBER 10, 2024:961–973
962
s a us is i ually comple e h ough ela ed con ac s
and in e mi en Na ional Dea h Index sea ches. The
CARDIA s udy was app o ed by he No hwes e n
Ins i u ional Re iew Boa d.
STUDY SAMPLE AND EXCLUSION CRITERIA. Fo he
p ima y analysis, eligible adul s we e 4,104 CARDIA
pa icipan s who had $1measu emen o as ing
lipids be ween ages 18 o 30 yea s, and a leas 1
measu emen be ween age 35 o 45 yea s. Indi iduals
who used lipid-lowe ing pha maco he apy be ween
ages 18 and 40 yea s we e excluded om his analysis
(n ¼88). Indi iduals who expe ienced an ASCVD
e en be o e age 40 yea s (n ¼21) we e excluded as
well. A o al o 3,995 pa icipan s we e included in he
final analysis se .
We pe o med a seconda y analysis o assess he
abili y o 2 measu emen s o a he ogenic lipids be o e
age 30 yea s o p edic he cumula i e exposu e
h ough age 40 yea s. O he 3,995 CARDIA pa ici-
pan s in he p ima y analysis, 1,806 pa icipan s
had $2measu emen so non–HDL-C o LDL-C p io
o age 30 yea s. These pa icipan s wi h $2mea-
su emen s we e included in he seconda y analysis.
TRADITIONAL RISK FACTOR ASSESSMENT. Age,
ace, and sex we e de e mined by sel - epo . Heigh ,
weigh , and wais ci cum e ence we e measu ed wi h
he pa icipan s in ligh clo hing using a s anda dized
s adiome e , ape measu e, and calib a ed scale; body
mass index (BMI) (kg/m
2
) was calcula ed. A each
CARDIA examina ion, pa icipan s we e gi en he
in e iewe -adminis e ed Physical Ac i i y His o y
Ques ionnai e.
9,10
The physical ac i i y sco e sum-
med equency imes in ensi y o e he 13 ac i i ies
o ge o al ac i i y (in exe cise uni s).
Smoking habi s and educa ional a ainmen
we e de e mined wi h he use o s anda dized
and alida ed ques ionnai es. Blood p essu e–and
choles e ol-lowe ing medica ion use was de e mined
by sel - epo .
Blood p essu e was measu ed a e 5 minu es o
es in he sea ed posi ion using a andom ze o me -
cu y sphygmomanome e , eplaced om Y20 o wa d
wi h an Om on oscillome e (calib a ed o he andom
ze o). The means o he second and hi d sys olic and
dias olic measu emen s we e used.
LABORATORY AND LIPID MEASUREMENTS. A e a
12-hou as , blood was d awn om a ein in he
an ecubi al ossa in o a Vacu aine , coa ed wi h
EDTA o plasma. Se um and plasma samples we e
ob ained and s o ed a –80 C o u u e analysis.
Samples we e shipped on d y ice o he No hwes
Lipid Resea ch Cen e in Sea le, Washing on.
Plasma concen a ions o o al choles e ol and
iglyce ides we e measu ed using a s anda d
enzyma ic assay. High-densi y lipop o ein choles-
e ol was quan ified a e p ecipi a ion wi h dex al
sul a e-magnesium chlo ide on an ABA 200 Bio-
ch oma ic ins umen (Abbo Labo a o ies).
8
To
educe e o in LDL-C es ima ion a high o low
LDL-C alues and in indi iduals wi h ele a ed i-
glyce ide le els, we used LDL-C es ima ion as pub-
lished by Sampson e al.
11
Non–HDL-C was
calcula ed as he di e ence be ween o al choles-
e ol and high-densi y lipop o ein choles e ol o
each pa icipan .
CALCULATION OF CUMULATIVE EXPOSURE. The
mean numbe o lipid measu emen s be ween ages
18 and 40 yea s was 5.1 (Q1-Q3: 5.0-6.0). Cumula i e
non–HDL-C and LDL-C we e defined based on he
p edic ed alue in each yea du ing ages 18 o
40 yea s. We chose an age-based analysis ins ead o
examina ion- ime-based analysis and we p esen
he isi ime a each examina ion by pa icipan s’
ages. We applied a nonpa ame ic qua ic splined-
based mixed model o es ima e he subjec -specific
non–HDL-C and LDL-C ajec o y om ages 18 o
40 yea s o all pa icipan s.
12
This app oach allows
aflexible shape o he indi idual ajec o ies and
es ima es he lipid le els o pa icipan s wi h
insu ficien da a by bo owing s eng h om o he
indi iduals. The a ea unde he cu e o lipid
exposu e om age 18 o 40 (AUC [18-40 yea s]) was
hen calcula ed as he o e all cumula i e exposu e
and exp essed in he uni o mg/dL yea s. This
me hod was selec ed due o i s abili y o ailo a
nonlinea lipid cu e o e he exposu e pe iod.
Cumula i e a he ogenic lipid alues we e hen
di ided in o qua iles. Fo ease o in e p e a ion, we
di ided he cumula i e alues by 22 yea s ( he fixed
ollow-up in e al) o ep esen usual yea ly non–
HDL-C o LDL-C.
13
ASSESSMENT OF ASCVD ENDPOINTS. We defined
ASCVD e en s as a al o non a al myoca dial in a c-
ion (MI) o s oke as adjudica ed by he CARDIA
Endpoin s Commi ee. We included all ASCVD e en s
ha occu ed a e pa icipan age 40 yea s and 2022.
Pa icipan s we e ollowed by yea ly elephone in-
e iews wi h ei he he pa icipan s o a designa ed
p oxy o assess in e im ASCVD e en s, hospi aliza-
ions, ou pa ien p ocedu es, o dea h. Medical e-
co ds we e eques ed o pa icipan s who had been
hospi alized o ecei ed an ou pa ien e ascula iza-
ion p ocedu e (eg, pe cu aneous co ona y in e en-
ion). Medical eco ds we e e iewed and adjudica ed
independen ly by 2 physician membe s o he end-
poin s commi ee. I disag eemen occu ed, cases
JACCVOL.84,NO.11,2024 Wilkins e al
SEPTEMBER 10, 2024:961–973 P edic ion o Lipid Exposu e in Young Adul s
963
we e e iewed by he en i e endpoin s commi ee o
ob ain consensus.
C i e ia o a al o non a al MI, and a al and
non a al s oke a e ou lined in he CARDIA manual o
ope a ions. B iefly, MI was defined as symp oms
consis en wi h ca diac ischemia, wi h ele a ions in
bioma ke s ( oponin, c ea inine kinase), wi h o
wi hou he p esence o consis en elec oca dio-
g aphic changes. S oke was defined as a neu ologic
defici ha las s>24 hou s and is consis en wi h
acu e dis up ion o blood flow o a ascula e i o y
wi h o wi hou suppo i e imaging findings on
compu ed omog aphy scan o magne ic esonance
imaging. A a al MI o ce eb o ascula acciden e en
was defined as dea h wi hin 30 days o ei he acu e
diagnosis.
STATISTICAL ANALYSIS. We c ea ed qua iles o
cumula i e exposu e o non–HDL-C o LDL-C sepa-
a ely and compa ed pa icipan cha ac e is ics,
including demog aphics, an h opome ics, li es yle
beha io s, and adi ional isk ac o s ac oss qua iles
using analysis o a iance o con inuous a iables
and chi-squa e es o ca ego ical a iables
as app op ia e.
We assessed he abili y o he fi s measu emen o
a he ogenic lipids be ween ages 18 and 30 yea s o
p edic ele a ed ( op qua ile: 75 h-100 h pe cen ile)
o low (0 h-25 h pe cen ile) cumula i e exposu e,
espec i ely, by age 40 yea s using ecei e -ope a ing
cha ac e is ic (ROC) cu es. Model 1 was a uni a ia e
logis ic eg ession model ha included a 1- ime
measu emen o non–HDL-C o LDL-C; Model 2
included Model 1 þage, sex, ace; Model 3 included
Model 2 þBMI and SBP. The compa ison o a eas
unde he ROC cu es gene a ed by hese models was
pe o med acco ding o DeLong e al.
14
We an he
analyses o non–HDL-C and LDL-C sepa a ely.
We calcula ed c oss- ables o he pe cen ages o
pa icipan s by he qua ile o hei fi s a he ogenic
lipid le el and he fi s measu emen made a e age
40 yea s. We also calcula ed c oss- ables o he pe -
cen ages o pa icipan s by he qua ile o hei fi s
a he ogenic lipid le el and he qua ile o cumula i e
exposu e h ough age 40 yea s. The ag eemen be-
ween he qua ile anking o hese 2 measu emen s
was assessed using he Cohen
k
s a is ic.
To define h esholds wi h op imal pe o mance o
disc imina e hosewhoa elikely oha eahigh s
low cumula i e exposu e, we used Youden’sindex o
de e mine he non–HDL-C and LDL-C absolu e alues
ha we e associa ed wi h he highes simul aneous
sensi i i y and specifici y o p edic membe ship in
he highes o lowes qua ile (sepa a ely) o cumu-
la i e exposu e.
The abili y o 2 measu emen s o a he ogenic lipids
be ween age 18 and 30 yea s o p edic membe ship in
he op qua ile o bo om qua ile o cumula i e
exposu e h ough age 40 yea s was assessed using
sepa a e models ha included 2 LDL-C o non–HDL-C
measu emen s ha we e made be o e age 30 yea s.
We anasecondse o models ha includedage,sex,
and ace co a ia es.
To quan i y associa ions be ween cumula i e
exposu e qua iles wi h long- e m isk o ASCVD
e en s, we used Cox p opo ional haza ds models o
each qua ile o non–HDL-C o LDL-C a fi s mea-
su emen (be ween ages 18 and 30 yea s) and o each
qua ile o cumula i e exposu e o non–HDL-C o
LDL-C (sepa a ely) om age 18 o 40 yea s wi h isk
o ASCVDe en sa e age40yea s.Modelswe e un
as ollows: Model 1: age-, sex-, and ace-adjus ed; and
Model 2: Model 1 þeduca ional a ainmen , BMI,
sys olic blood p essu e, blood p essu e ea men ,
diabe es, diabe es ea men , and cu en obacco
use. Fo qua iles o cumula i e exposu e, we
included a hi d model ha included all co a ia es
om Model 2 plus he alue o he fi s non–HDL-C o
LDL-C ha was made be ween ages 18 and 30 yea s.
We used co a ia es in Model 2 om he examina ion
ha was closes o when he pa icipan was age
40 yea s.
RESULTS
BASELINE CHARACTERISTICS BY QUARTILES OF
NON–HDL-C CUMULATIVE EXPOSURE. Cha ac e is-
ics o CARDIA pa icipan s a ime o he fi s lipid
measu emen , s a ified by qua ile o cumula i e
non–HDL-C and LDL-C exposu e om age 18 o
40 yea s, a e shown in Tables 1 and 2, espec i ely.
The mean age was 25 yea s a he ime o fi s non–
HDL-C measu emen . The mean non–HDL-C was
app oxima ely 10 mg/dL highe han LDL-C in young
adul s. The mean non–HDL-C a age 25 yea s in he
highes non–HDL-C qua ile was 161.9 mg/dL, and he
mean LDL-C a age 25 yea s in he highes LDL-C
qua ile was 146.1 mg/dL. Pa icipan s in he highe
qua iles o non–HDL-Cexposu ewe emo elikely o
be men and iden i y as Whi e. Mean baseline BMI,
sys olic and dias olic blood p essu es, and as ing
glucosewe esligh lyhighe inhighe compa edwi h
lowe qua iles o non–HDL-C cumula i e exposu e.
No ably, in his young adul sample, all he mean
alues o hese measu es, excep o BMI, we e wi hin
anges conside ed clinically “no mal” o middle-aged
adul s. Simila dis ibu ions o baseline cha ac e is ics
we e obse ed ac oss qua iles o LDL-C as o
non–HDL-C (Table 2).
Wilkins e al JACC VOL. 84, NO. 11, 2024
P edic ion o Lipid Exposu e in Young Adul s SEPTEMBER 10, 2024:961–973
964
Cha ac e is ics o CARDIA pa icipan s a he ex-
amina ion closes o age 40 yea s by qua ile o cu-
mula i e non–HDL-C and LDL-C exposu e ( om ages
18 o 40 yea s) a e shown in Supplemen al Tables 1A
and 1B, espec i ely. The mean non–HDL-C was
abou 20 mg/dL highe han he LDL-C a age 40
yea s. Dis ibu ions o o he cha ac e is ics ac oss
qua iles we e simila o hose obse ed a baseline,
al hough absolu e alues o all con inuous isk ac o s
we e quali a i ely highe han a baseline.
DISTRIBUTION OF FIRST NON-HDL-C OR LDL-C
MEASUREMENTS BY CUMULATIVE EXPOSURE
QUARTILE. The dis ibu ions o fi s measu emen o
non–HDL-C and LDL-C, by qua ile o cumula i e
exposu e be ween ages 18 and 40 yea s, a e shown in
Figu e 1A o non–HDL-C and Figu e 1B o LDL-C.
No ably, he mean fi s non–HDL-C measu ed a
mean age 25 yea s di e ed ac oss cumula i e non–
HDL-C exposu e qua iles, ye o e lap in he alues o
he fi s measu emen is p esen o all 4 qua iles,
demons a ing eclassifica ion in qua ile o ini ial
measu emen be o e age 30 yea s and qua ile o cu-
mula i e exposu e by age 40 yea s. Ye , he e is
limi ed o e lap in baseline measu emen s be ween
he highes and lowes qua iles o cumula i e expo-
su e, demons a ing ha a modes pe cen age o
indi iduals who we e in he middle o he dis ibu-
ion a hei ini ial measu emen had inc eases o
dec eases in lipid le els ha pu hem in he highes
o lowes qua ile o cumula i e exposu e by age
40 yea s.
QUARTILE OF FIRST NON–HDL-C MEASUREMENT BY
QUARTILE OF FIRST MEASUREMENT MADE AFTER
AGE 40 YEARS. C oss- ables displaying he qua ile
o he fi s non–HDL-C measu emen aken p io o
age 30 yea s and he fi s measu emen a e age 40
yea s a e p esen ed in Supplemen al Table 2A.O he
pa icipan s ini ially in he op and bo om qua iles
based on hei fi s measu emen , 57.7% emained in
hesamequa ilea e age40yea s.Con e sely,3.9%
o hose ini ially in he op qua ile shi ed o he
bo om qua ile a e age 40 yea s, whe eas 3.1% o
hose ini ially in he bo om qua ile mo ed o he op
qua ile a e age 40 yea s. Simila esul s we e
obse ed o LDL-C.
QUARTILE OF FIRST NON–HDL-C AND LDL-C AND
QUARTILE OF CUMULATIVE EXPOSURE. Ac oss-
able displaying he qua ile o he fi s non–HDL-C
akenbe o eage30yea sand hequa ileo cumu-
la i e exposu e is shown in Supplemen al Table 2B.
App oxima ely 75% o pa icipan s who s a ed in he
op o bo om qua ile o non–HDL-C was in he same
TABLE 1 Cha ac e is ics o CARDIA Pa icipan s a he Age o Fi s Non–HDL-C Measu emen
Qua ile o Cumula i e Non–HDL-C Exposu e
(Range o Cumula i e Exposu e, mg/dL y)
0%-25%
(n ¼998)
(622-2,415)
25%-50%
(n ¼999)
(2,416-2,817)
50%-75%
(n ¼999)
(2,819-3,243)
75%-100%
(n ¼999)
(3,244-6,117)
Usual
a
yea ly non–HDL-C om age 18-40 y, mg/dL/y 98 (28-110) 119 (111-128) 137 (129-147) 162 (148-278)
Non–HDL-C, mg/dL 88.9 17.1 113.2 17.2 129.3 17.5 161.9 27.1
Age, y 24.9 3.6 24.9 3.6 25.3 3.5 25.4 3.5
Black 503 (50.4) 495 (49.5) 477 (47.7) 461 (46.1)
Male 358 (35.9) 372 (37.2) 458 (45.8) 569 (57.0)
Educa ion, y 15.8 2.7 15.8 4.5 15.6 2.5 15.5 3.7
To al choles e ol, mg/dL 147.2 20.7 167.7 21.3 181.5 20.9 210.4 28.5
HDL-C, mg/dL 58.2 13.3 54.6 12.3 52.3 12.7 48.6 11.6
LDL-C, mg/dL 78.1 17.0 100.4 16.8 114.6 17.7 143.2 25.8
Body mass index, kg/m
2
23.2 4.4 24.3 5.0 24.8 5.1 25.7 5.1
Sys olic blood p essu e, mm Hg 105.7 11.0 107.0 10.8 107.9 11.7 109.8 12.0
Dias olic blood p essu e, mm Hg 67.3 9.9 68.5 9.7 69.6 9.8 71.1 10.5
Fas ing glucose, mg/dL 80.8 8.9 81.4 8.6 82.2 11.3 83.2 10.4
Cu en smoke 282 (28.3) 265 (26.6) 275 (27.7) 299 (30.2)
Physical ac i i y in ensi y sco e 433.2 293.1 421.4 307.3 403.3 292.3 417.3 302.7
An ihype ensi e medica ion use 16 (1.6) 18 (1.8) 20 (2.0) 30 (3.0)
Hypoglycemic medica ion use 3 (0.3) 3 (0.3) 8 (0.8) 4 (0.4)
Diabe es 3 (0.3) 3 (0.3) 2 (0.2) 2 (0.2)
ASCVD e en s a e age 40 y 21 (2.1) 57 (5.7) 59 (5.9) 106 (10.6)
Values a e median ( ange), mean SD, o n (%).
a
Usual non–high-densi y lipop o ein choles e ol (non–HDL-C) ¼cumula i e non–HDL-C/22 yea s.
ASCVD ¼a he oscle o ic ca dio ascula disease; HDL-C ¼high-densi y lipop o ein choles e ol; LDL-C ¼low-densi y lipop o ein choles e ol.
JACCVOL.84,NO.11,2024 Wilkins e al
SEPTEMBER 10, 2024:961–973 P edic ion o Lipid Exposu e in Young Adul s
965
qua ile o cumula i e exposu e hough age 40 yea s.
The ag eemen be ween he qua ile ankings o hese
2 measu emen s we e 62.8% (kappa ¼0.66) o non–
HDL-C. Simila esul s we e obse ed o LDL-C.
UNIVARIABLE AND MULTIVARIABLE MODELS TO
PREDICT CUMULATIVE NON–HDL-C AND LDL-C. The
ROC cu es o models o p edic cumula i e non–
HDL-C exposu e a e shown in Figu es 2A and 2B.Fo
p edic ing membe ship in he highes qua ile o cu-
mula i e exposu e, he C-s a is ics o 1- ime baseline
measu es o non–HDL-C and LDL-C we e bo h 0.93.
Addi ional adjus men o age, sex, ace, BMI, and
sys olic blood p essu e (measu es commonly ob-
ained du ing clinical isi s) did no significan ly
change he c-s a is ic.
The ROC cu es o models o p edic membe ship
in helowes qua ileo cumula i eexposu e onon–
HDL-C and LDL-C a e shown in Figu es 2A and 2B.The
c-s a is ics o a 1- ime measu e be o e age 30 yea s o
p edic membe ship in he lowes qua ile o cumu-
la i e exposu e we e also 0.93 o non–HDL-C and
LDL-C, espec i ely.Addi iono age,sex, ace,BMI,
and sys olic blood p essu e did no significan ly
change he C-s a is ics.
In seconda y analysis, 2 measu emen s o non–
HDL-C o LDL-C be o e age 30 yea s p edic ed mem-
be shipin he opo bo omqua ileo cumula i e
exposu e, wi h a c-s a is ic o 0.97 and 0.97,
espec i ely.
COMPARISONS OF ABSOLUTE LIPID VALUES THAT
PREDICT MEMBERSHIP IN THE HIGHEST AND
LOWEST QUARTILES OF CUMULATIVE EXPOSURE.
The pe o mance o he absolu e lipid alues wi h he
highes simul aneous sensi i i y and specifici y
(highes Youden’s index) om ou model o p edic
membe ship in he highes qua ile o cumula i e
exposu e a e shown in Table 3.Fo non–HDL-C, a
1- ime alue >135 mg/dL be ween ages 18 and
30 yea s had 85% sensi i i y and 86% specifici y o
p edic membe ship in he op qua ile o cumula i e
non–HDL-C exposu e by age 40 yea s. In compa ison,
anon–HDL-C alue o >160 mg/dL be ween ages 18
and 30 yea s, which is commonly used in clinical
p ac ice, had sensi i i y o 48%, bu highe specifici y
o 98% o p edic membe ship in he op qua ile o
cumula i e exposu e o non–HDL-C by age 40 yea s.
Fo LDL-C, a alue >118 mg/dL be ween ages 18 and
30 yea s had sensi i i y o 89% and specifici y o 82%
TABLE 2 Cha ac e is ics o CARDIA Pa icipan s a he Age o Fi s LDL-C Measu emen
Qua ile o Cumula i e LDL-C Exposu e
(Range o Cumula i e Exposu e, mg/dL y)
0%-25%
(n ¼998)
(327-2,106)
25%-50%
(n ¼999)
(2,107-2,485)
50%-75%
(n ¼999)
(2,486-2,885)
75%-100%
(n ¼999)
(2,886-5,905)
Usual
a
yea ly non–HDL-C om age 18-40 y, mg/dL/y 85 (15-96) 105 (97-113) 122 (114-131) 144 (132-268)
LDL-C, mg/dL 76.8 17.0 99.5 15.3 116.3 16.8 146.1 25.0
Age, y 24.9 3.6 24.9 3.6 25.2 3.6 25.6 3.4
Black 497 (49.8) 468 (46.8) 480 (48.0) 491 (49.1)
Male 383 (38.4) 398 (39.8) 441 (44.1) 535 (53.6)
Educa ion, y 15.7 2.7 15.8 4.5 15.7 2.6 15.5 3.7
To al choles e ol, mg/dL 146.3 20.5 166.5 19.4 182.6 20.4 211.4 27.6
HDL-C, mg/dL 57.2 13.8 53.9 12.4 52.4 12.7 50.1 11.8
LDL-C, mg/dL 89.1 17.8 112.6 16.5 130.2 18.6 161.3 27.0
Body mass index, kg/m
2
23.4 4.5 24.4 5.0 24.5 5.0 25.6 5.1
Sys olic blood p essu e, mm Hg 106.3 11.4 106.8 10.6 108.0 11.5 109.2 12.1
Dias olic blood p essu e, mm Hg 67.7 10.1 68.5 9.7 69.7 9.8 70.5 10.5
Fas ing glucose, mg/dL 81.0 8.9 81.5 8.8 82.2 11.3 82.8 10.3
Cu en smoke 304 (30.5) 264 (26.5) 263 (26.4) 290 (29.3)
Physical ac i i y in ensi y sco e 435.9 290.0 427.9 316.8 395.8 285.0 415.7 302.1
An ihype ensi e medica ion use 17 (1.7) 17 (1.7) 24 (2.4) 26 (2.6)
Hypoglycemic medica ion use 3 (0.3) 3 (0.3) 3 (0.3) 1 (0.1)
Diabe es 3 (0.3) 3 (0.3) 8 (0.8) 4 (0.4)
ASCVD e en s a e age 40 y 22 (2.2) 61 (6.1) 62 (6.2) 98 (9.8)
Values a e median ( ange), mean SD, o n (%).
a
Usual non–HDL-C ¼cumula i e non–HDL-C/22 yea s.
Abb e ia ions as in Table 1.
Wilkins e al JACC VOL. 84, NO. 11, 2024
P edic ion o Lipid Exposu e in Young Adul s SEPTEMBER 10, 2024:961–973
966
o p edic membe ship in he op qua ile o cumu-
la i e LDL-C exposu e by age 40 yea s. A commonly
used LDL-C h eshold >130 mg/dL be ween ages 18
and30yea shadsensi i i yo 72%andspecifici y o
93% o membe ship in he op qua ile o cumula i e
LDL-C exposu e by age 40 yea s.
The pe o mance o he absolu e lipid alues wi h
he highes simul aneous specifici y and sensi i i y o
p edic membe ship in he lowes qua ile o cumu-
la i e exposu e, compa ed wi h alues commonly
used in clinical p ac ice, a e shown in Table 4.Anon–
HDL-C alue o <107 mg/dL be ween ages 18 and 30
yea s had a specifici y o 86% and a sensi i i y o 87%
o membe ship in he lowes qua ile o cumula i e
non–HDL-C exposu e by age 40 yea s, whe eas he
commonly used alue o <130 mg/dL be ween ages 18
and30yea shadaspecifici y o 51% and a sensi i i y
o 99%. An LDL-C alue o <96 mg/dL be ween ages
18and30yea shadaspecifici y o 84% and a sensi-
i i y o 89% o membe ship in he bo om qua ile
o cumula i e exposu e by age 40 yea s. These alues
we e simila o hose ob ained when he commonly
used LDL-C alue o <100mg/dLwasused o young
adul s.
The highes specifici y o membe ship in he op
o bo om qua iles can be de e mined om isual in-
spec ion o Figu e 1 and confi med in Figu es 2A and 2B.
All indi iduals wi h non–HDL-C alues below
60 mg/dL we e in he bo om qua ile o cumula i e
exposu e, whe eas all pa icipan s wi h an ini ial
measu emen o 180 mg/dL o g ea e we e in he op
qua ile o cumula i e exposu e.
ASCVD RISKS BY QUARTILE OF NON–HDL-C AND
LDL-C. A e age 40 yea s, he 3,995 CARDIA pa ici-
pan s con ibu ed o a o al o 77,141 pe son-yea s o
ollow-up, wi h a mean ollow-up ime o 19.2 5.4
yea s.The a eo ASCVDe en sa e heageo 40
yea s was 3.15 pe 1,000 pe son-yea s. The Kaplan
Meie su i al cu es o he su i al p obabili y ee
om ASCVD a e age 40 yea s by qua iles o
cumula i e non-HDL-C and LDL-C exposu e a e p e-
sen ed in Supplemen al Figu es 1A and 1B, espec-
i ely. The HRs o ASCVD e en s a e age 40 yea s
by qua ile o non–HDL-C and LDL-C exposu e om
ages 18 o 40 yea s a e shown in Table 5.One- ime
measu emen s o non–HDL-C and LDL-C ob ained
be ween ages 18 and 30 yea s we e posi i ely asso-
cia ed wi h ASCVD isk a e age 40 yea s, wi h a isk-
FIGURE 1 Densi y Plo : Fi s Non–HDL-C o LDL-C by Exposu e Qua iles
0.000
0.005
0.010
0.015
0.020
0 50 100
Fi s Non-HDL-C (mg/dL)
AB
Densi y
150 200 250
Cumula i e Non-HDL-C Qua ile
0%-25% 25%-50%
50%-75% 75%-100%
0.000
0.005
0.010
0.015
0.025
0.020
050 100
Fi s LDL-C (mg/dL)
Densi y
150 200 250
Cumula i e LDL-C Qua ile
0%-25% 25%-50%
50%-75% 75%-100%
Densi y plo o he fi s non–high-densi y lipop o ein choles e ol (non–HDL-C) (A) o low-densi y lipop o ein choles e ol (LDL-C) (B) mea-
su emen made be ween ages 18 and 30 yea s in he CARDIA (Co ona y A e y Risk De elopmen in Young Adul s S udy) coho and he
qua ile o cumula i e exposu e om age 18 h ough 40 yea s. The blue cu e ep esen s he lowes qua ile, ed and g ay a e sequen ially
highe qua iles, and pu ple is he highes qua ile o cumula i e exposu e. O no e, o e lap in he alues o he fi s measu emen s a e
p esen o all 4 qua iles. Thus, subs an ial eclassifica ion be ween qua ile o fi s measu emen and cumula i e exposu e in ea ly adul
li e occu s. Howe e , he e is modes o e lap be ween he op and bo om qua iles, demons a ing ha a modes pe cen age o indi iduals
who we e in he middle o he dis ibu ion a hei ini ial measu emen had inc eases o dec eases in lipid le els ha pu hem in he highes o
lowes qua ile o cumula i e exposu e by age 40 yea s.
JACCVOL.84,NO.11,2024 Wilkins e al
SEPTEMBER 10, 2024:961–973 P edic ion o Lipid Exposu e in Young Adul s
967
FIGURE 2 ROC Cu es o he Cumula i e Exposu e P edic ion Models
ROC Cu es o Compa isons ROC Cu es o Compa isons
A
B
0.00
0.25
0.50
0.75
1.00
0.00 0.25 0.50 0.75
1.00
1- Speci ici y
Non-HDL-C Top Qua ile
ROC Cu es o Compa isons
Sensi i i y
0.00
0.25
0.50
0.75
1.00
0.00 0.25 0.50 0.75
1.00
1- Speci ici y
LDL-C Top Qua ile
Sensi i i y
0.00
0.25
0.50
0.75
1.00
0.00 0.25 0.50 0.75 1.00
1- Speci ici y
Non-HDL-C Bo om Qua ile
ROC Cu es o Compa isons
Sensi i i y
0.00
0.25
0.50
0.75
1.00
0.00 0.25 0.50 0.75 1.00
1- Speci ici y
LDL-C Bo om Qua ile
Sensi i i y
ROC Cu e (A ea)
Unadjus ed (0.9304)
Age/sex/ ace (0.9368)
Age/sex/ ace/BMI + SBP
(0.9369)
ROC Cu e (A ea)
Unadjus ed (0.9325)
Age/sex/ ace (0.9355)
Age/sex/ ace/BMI + SBP
(0.9355)
ROC Cu e (A ea)
Unadjus ed (0.9332)
Age/sex/ ace (0.9352)
Age/sex/ ace/BMI + SBP
(0.9353)
ROC Cu e (A ea)
Unadjus ed (0.9338)
Age/sex/ ace (0.9372)
Age/sex/ ace/BMI + SBP
(0.9375)
ROC cu es o he p edic ion o membe ship in he op o bo om qua ile o cumula i e exposu e o (A) non–HDL-C and (B) LDL-C om age 18
o 40 yea s by a 1- ime measu emen made be ween ages 18 and 30 yea s. O e all disc imina o y capaci y is high (c-s a is ics o 0.93 o all
models). The addi ion o demog aphic a iables, body mass index (BMI), and sys olic blood p essu e (SBP) yielded e y modes e ec s on he
C-s a is ics. ROC ¼ ecei e -ope a ing cha ac e is ic; o he abb e ia ions as in Figu e 1.
Wilkins e al JACC VOL. 84, NO. 11, 2024
P edic ion o Lipid Exposu e in Young Adul s SEPTEMBER 10, 2024:961–973
968
ac o adjus ed HR o ASCVD e en s o 3.7 (95% CI:
2.3-6.1) and 2.5 (95% CI: 1.6-3.8) o he highes
qua iles o fi s non–HDL-C and LDL-C measu emen ,
espec i ely,whencompa edwi h helowes qua -
iles. When compa ed wi h he lowes qua ile, hose
in he highes qua ile o cumula i e non–HDL-C and
LDL-C had adjus ed HRs o 4.6 (95% CI: 2.8-7.3) and
4.0 (95% CI: 2.5-6.3) espec i ely, o ASCVD e en s
a e age 40 yea s. The s eng h o associa ion was
modes ly educed by mul i a iable adjus men o
o he adi ional isk ac o s (Model 2) including
adjus men o he fi s lipid measu emen made be-
ween ages 18-30 yea s (Model 3).
DISCUSSION
In his longi udinal coho o young adul s wi h
epea ed lipid measu es, we obse ed ha 1- ime
measu emen o non–HDL-C o LDL-C ob ained be-
ween ages 18 and 30 yea s can p edic membe ship
in he op and bo om qua iles o cumula i e expo-
su e o a he ogenic lipids om age 18 o 40 yea s wi h
a e y high deg ee o disc imina ion. The age pe iod
be ween 18 o 40 yea s is a c i ical ime o exposu e o
CVD isk de e minan s including a he ogenic lipids.
2
Thus, iden i ying indi iduals likely o ha e he
highes (o lowes ) cumula i e exposu es wi h a
1- ime measu emen may be o significan clinical
alue o isk communica ion and p e en i e
decision-making in young adul s.
Absolu e alues o non–HDL-C >135 mg/dL o
LDL-C >118 mg/dL be ween ages 18 and 30 yea s we e
highly sensi i e and specific o iden i ying young
adul s who we e likely o ha e a high cumula i e
exposu e o hose espec i e a he ogenic lipid ac-
ions (Cen al Illus a ion). No ably, hese h esholds
a e conside ably lowe han alues commonly used in
clinical p ac ice (ie, non–HDL-C >160 mg/dL and
LDL-C >130 mg/dL) o indica e “ele a ion.”
Con e sely, non–HDL-C alues <107 mg/dL had sub-
s an ially highe specifici y han he cu en ly used
alue o <130 mg/dL o p edic membe ship in he
bo om qua ile o exposu e. Absolu e alues o
LDL-C o <96 mg/dL, e y close o he commonly
used h eshold o <100 mg/dL, p edic ed membe ship
in he bo om qua ile wi h a specifici y o 84% and a
high sensi i i y.
As expec ed, membe ship in he highes qua ile o
1- ime measu emen on non–HDL-C and LDL-C ob-
ained be ween ages 18 and 30 yea s and he cumu-
la i e exposu e om age 18 o 40 yea s was s ongly
associa ed wi h ASCVD e en isk a e age 40 yea s
when compa ed wi h he lowes qua ile o cumula-
i e exposu e. The associa ions be ween qua ile o
cumula i e exposu e and ASCVD e en s a e age
40 yea s we e obus o adjus men o he fi s
TABLE 3 Tes Cha ac e is ics o Th esholds o P edic Membe ship in he Highes Qua ile o Cumula i eExposu ebyAge40Yea s
Th eshold, mg/dL
Sensi i i y
%(95%CI)
Specifici y
%(95%CI)
NPV
%(95%CI)
PPV
%(95%CI)
Accu acy
%(95%CI)
Non–HDL-C
>135
a
84.5 (82.3-86.8) 86.2 (84.9-87.5) 94.3 (93.5-95.2) 67.2 (64.6-69.8) 85.8 (84.7-86.9)
>160 48.4 (44.9-51.5) 98.2 (97.7-98.7) 85.0 (83.8-86.2) 90.1 (87.5-92.6) 85.7 (84.6-86.7)
LDL-C
>118
a
88.8 (86.8-90.8) 82.0 (80.6-83.4) 95.6 (94.8-96.4) 62.2 (60.0-64.7) 83.7 (82.5-84.8)
>130 71.7 (68.9-74.5) 93.3 (92.4-94.2) 90.8 (89.7-91.8) 78.1 (75.4-80.8) 87.7 (86.8-88.9)
a
Value de i ed om ecei e -ope a ing cha ac e is ic cu e analysis (highes Youden’s Index).
NPV ¼nega i e p edic i e alue; PPV ¼posi i e p edic i e alue; o he abb e ia ions as in Table 1.
TABLE 4 Tes Cha ac e is ics o Th esholds o P edic Membe ship in he Lowes Qua ile o Cumula i e Exposu e by Age 40 Yea s
Th eshold, mg/dL
Sensi i i y
%(95%CI)
Specifici y
%(95%CI)
NPV
%(95%CI)
PPV
%(95%CI)
Accu acy
%(95%CI)
Non–HDL-C
<107
a
86.8 (84.7-88.8) 85.7 (84.5-87.0) 95.1 (94.3-95.9) 67.0 (64.4-69.5) 86.0 (84.9-87.1)
<130 99.1 (98.5-99.7) 50.9 (49.1-52.7) 99.4 (99.0-99.8) 40.2 (38.3-42.1) 62.9 (61.5-64.5)
LDL-C
<96
a
88.6 (86.6-90.6) 83.5 (82.1-84.8) 95.6 (94.8-96.4) 64.1 (61.5-66.6) 84.7 (83.6-85.8)
<100 93.2 (91.6-94.7) 78.2 (76.7-79.7) 97.2 (96.6-97.8) 58.7 (56.3-61.6) 81.9 (80.8-83.1)
a
Value de i ed om ecei e -ope a ing cha ac e is ic cu e analysis (highes Youden’s Index).
Abb e ia ions as in Tables 1 and 3.
JACCVOL.84,NO.11,2024 Wilkins e al
SEPTEMBER 10, 2024:961–973 P edic ion o Lipid Exposu e in Young Adul s
969
