prognostic value of fasting vs. non-fasting low density...
Post on 30-Apr-2020
2 Views
Preview:
TRANSCRIPT
BangaloreBethany Doran, Yu Guo, Jinfeng Xu, Howard Weintraub, Samia Mora, David J. Maron and Sripal
(NHANES-III)Long-term Mortality: Insight from the National Health and Nutrition Survey III Prognostic Value of Fasting vs. Non-Fasting Low Density Lipoprotein Cholesterol Levels on
Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 2014 American Heart Association, Inc. All rights reserved.
is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation published online July 11, 2014;Circulation.
http://circ.ahajournals.org/content/early/2014/07/10/CIRCULATIONAHA.114.010001World Wide Web at:
The online version of this article, along with updated information and services, is located on the
http://circ.ahajournals.org/circulationaha/suppl/2014/07/10/CIRCULATIONAHA.114.010001.DC1.htmlData Supplement (unedited) at:
http://circ.ahajournals.org//subscriptions/
is online at: Circulation Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer available in the
Permissions in the middle column of the Web page under Services. Further information about this process isOnce the online version of the published article for which permission is being requested is located, click Request
can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Circulation Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
1
Prognostic Value of Fasting vs. Non-Fasting Low Density Lipoprotein
Cholesterol Levels on Long-term Mortality: Insight from the National Health
and Nutrition Survey III (NHANES-III)
Running title: Doran et al.; Fasting vs. non-fasting lipids
Bethany Doran, MD, MPH1; Yu Guo, MA1; Jinfeng Xu, PhD1; Howard Weintraub, MD1;
Samia Mora, MD2; David J. Maron, MD3; Sripal Bangalore, MD, MHA1
1New York University School of Medicine, New York, NY; 2Brigham and Women’s Hospital,
Boston, MA; 3Stanford University, Stanford, CA
Address for Correspondence:
Sripal Bangalore, MD, MHA, FACC, FAHA, FSCAI
Director of Research, Cardiac Catheterization Laboratory
Director, Cardiovascular Outcomes Group
Associate Professor of Medicine, New York University School of Medicine
New York, NY 10016
Tel: 212-263-3540
Fax: 212-263-3988
E-mail: sripalbangalore@gmail.com
Journal Subject Code: Atherosclerosis:[90] Lipid and lipoprotein metabolism
Samia Mora, MD2; David J. Maron, MD3; Sripal Bangalore, MD, MHMHHAAA1
11NeNeNeww w YYoYorkrkrk UUniniivveversr ity School of Medicine, Neweww YYYork, NY; 2Brrrigii haam m m aand Women’s Hospital,
BoBoststtoonon, MAMAMA; 33SSttanffforrrd UUUnniiveersrsrsitity,y,y, SSStaannnfooord,d CCCAAA
Address forr r CoCoCorrrrrresesespopopondndn enenenccec ::
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
2
Abstract
Background—National and international guidelines recommend fasting lipid panel measurement
for risk stratification of patients for prevention of cardiovascular (CV) events. Yet, the
prognostic value of fasting vs. non-fasting low density lipoprotein cholesterol (LDL-C) is
uncertain.
Methods and Results—Patients enrolled in the National Health and Nutrition Survey III
(NHANES-III), a nationally representative cross-sectional survey performed between 1988 to
1994, were stratified based on fasting status ( 8 hours or <8 hours) and followed for a mean of
14.0 (±0.22) years. Propensity score matching was used to assemble fasting and non-fasting
cohorts with similar baseline characteristics. The risk of outcomes as a function of LDL-C and
fasting status was assessed using receiver operating characteristic (ROC) curves and
bootstrapping methods. The interaction between fasting status and LDL-C was assessed using
Cox proportional hazards modeling. Primary outcome was all-cause mortality. Secondary
outcome was CV mortality. One-to-one matching based on propensity score yielded 4,299 pairs
of fasting and non-fasting individuals. For the primary outcome, fasting LDL-C yielded similar
prognostic value as non-fasting LDL-C [C-statistics=0.59 (95% CI 0.57-0.61) vs. 0.58 (95% CI
0.56-0.60; P=0.73], and LDL-C by fasting status interaction term in the Cox proportional hazard
model was not significant (Pinteraction= 0.11). Similar results were seen for the secondary outcome
[fasting vs. non-fasting C-statistics=0.62 (95% CI 0.60-0.66) vs. 0.62 (95% CI 0.60-0.66);
P=0.96; and Pinteraction=0.34].
Conclusions—Non-fasting LDL-C has similar prognostic value as that of fasting LDL-C.
National and international agencies should consider re-evaluating the recommendation that
patients fast before obtaining a lipid panel.
Key Words: cholesterol, mortality, fasting
fasting status was assessed using receiver operating characteristic (ROC) curvess ananand d d
bootstrapping methods. The interaction between fasting status and LDL-C was asasseseesssssededd uusissingnng
Cox proportional hazards modeling. Primary outcome was all-cause mortality. Secondary
ouutctccomommee wawawas CVVV mmmortality. One-to-one matchingngng bbased on propenennsityty ssscccoro e yielded 4,299 pairs
ofoff fffasasting andndd nnnoonon-f-fasasastititingngng iindndndivivividididuauaalslss.. FFFororr ttthhhee e prprimimimaryy y ouououtctct omomome,e,e fffasasastititinngng LDLDLDL-C-CC yyyieiei ldlddededed sssimimimilililar
prprrogggnon stic vvalaluee aass nooon--faststtininng g g LDLDL-LL CCC [[C-ssstaaatistiiicss=0=00.5.5.59 99 (9(( 555% CCII 0.557--0-0.6661)1)) vvs. 00.5588 (9995%%% CCCI
0.0 56566-0-0-0.6.660;0;0; PPP=0=0.7.773]3]3], , anananddd LDLDL L-L-L CCC bybyby ffasastititingngng sstatatatututusss inininteteterracacactititiononon ttteeerm mm ininin tthehehe CCCoxox ppprororopopoportrtioioonananal l l hahahazazazardrdr
model was nonoott t sisis gngngnififi iciccanana t (P(P(Pintinin erararactictict ononn== 0.0.0.11111).).. SiSiSimimimilalaar r r rer sususultltts ss wewewererere seeeeeen n n fofoforr r thththee sesesecococondndndara y outcomee
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
3
Introduction
Current national and international guidelines on cholesterol management recommend that lipid
panel measurement should be performed after an 8-12 hour fast.1-3 The reason often stated for
obtaining a fasting lipid panel is for greater precision for certain lipid parameters (especially
triglycerides), which can be variable, based on time and content of the last meal. From a practical
standpoint, it is cumbersome for patients to fast before obtaining a blood draw, and may delay
diagnosis and treatment of hyperlipidemia.
Prior data have shown that levels of total cholesterol (TC), low density lipoprotein
cholesterol (LDL-C), and high density lipoprotein cholesterol (HDL-C) vary little with respect to
fasting time, while triglycerides may vary by up to 20-30%.4, 5 Recently, studies have suggested
that non-fasting lipids may be equivalent (and potentially superior) in predicting cardiovascular
(CV) outcomes, as the non-fasting state may more accurately reflect the body’s exposure to
circulating lipids.6-8 Studies have demonstrated no benefit, or even improved risk prediction,
with the use of non-fasting as compared to fasting triglycerides.9-12 No prior studies have
examined the relationship of fasting vs. non-fasting LDL-C and mortality. Our objective was to
use the National Health Examination and Nutrition Survey III (NHANES-III), a nationally
representative database of the US population, to evaluate the prognostic value of fasting versus
non-fasting LDL-C for prediction of all-cause mortality and CV mortality in men and women.
Methods
Study Population
We used the NHANES-III linked to the National Death Index (NDI), a nationally representative
civilian cohort of non-institutionalized individuals within the United States. Baseline data was
fasting time, while triglycerides may vary by up to 20-30%.4, 5 Recently, studiess hhhavvve susuggggggesesesteted
hat non-fasting lipids may be equivalent (and potentially superior) in predicting cardiovascular
CCV)V)V) oooutututcococommemes, aaasss tht e non-fasting state may mooorerer aaccurately reflflecee t ththhee e bbody’s exposure to
ciircccuulating lipipidsdsds.66--8 SSSttudididieseses hhhaavave e dededemmoonnnstraaateed nooo bbbennenefffitit,, orrr eevvennn imimppproovovededed rrisisk kk prprredededici tititiononn,, ffff
wiwiiththth tttheheh uusesese oof f nnnonn-n-fafaaststiiningg asass ccooompmpmparareeded tttooo fffastststinini ggg ttrrigigglylylycececerriridededes.s.s 99-19-122 NNNooo prprrioior rr ststtududu ieieiess s haaaveee
examined thee rrrelele atatatioioionsnsshihih p p ofofof fffassstititingngng vvs.s.s. nnnononon-f-ffasasasttinining gg LDLDL L-L-L CCC ananand d d momomortrtrtalala ititity.y.y. OOOururur ooobjbjb ececectitive was to
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
4
collected between 1988-1994 using a multistage stratified probability cluster sampling design
where certain groups were intentionally oversampled and participant weights were added to
reflect the demographics of the 1990 US census. Comprehensive data about the validation and
collection of data are available elsewhere.13 The inclusion criteria for this study were adults 18
years of age and older residing in the United States who had participated in the NHANES III
study with data on fasting time. We excluded those in whom LDL-C calculations was not
possible due to missing HDL-C, TC, or triglyceride levels and those with triglycerides 400
mg/dl in whom the Friedewald equation may not be accurate.
Data Collection
Participants were interviewed in their homes and examined in a mobile examination center
where blood samples were obtained and physical exam performed. If participants were unable to
attend an examination at a center, home exam was performed. Institutional Review Board (IRB)
approval and documented consent was obtained from individuals through the Centers for Disease
Control and Prevention.
Laboratory Methods
Blood samples were collected through venipuncture and shipped on dry ice to the laboratory
analyzing the sample. Serum HDL-C, triglyceride, and TC levels were measured enzymatically
at Johns Hopkins University Lipoprotein Analytical Laboratory using a Hitachi 704 Analyzer
(Boehringer Mannheim Diagnostics, Indianapolis, Indiana). Lipid collection and analyses were
standardized to Centers for Disease Control and Prevention criterion.14 LDL-C was derived using
the Friedewald formula [LDL-C=TC - HDL-C – (triglycerides/5)],15 with prior studies showing
excellent correlation between fasting direct and indirect methods of LDL-C measurement,16-18
and a 0.97 correlation coefficient between Friedewald and directly measured LDL-C in non-
Participants were interviewed in their homes and examined in a mobile examinatattioiion nn cceentntntererer
where blood samples were obtained and physical exam performed. If participants were unable tom
attteteendndnd aaann exexexaamminnnatatatioion at a center, home exam wwwaaas s pperformed. Innstss itututtioioionnnal Review Board (IRB)
appprprovo al and ddococcuumumenennteeeddd cococonsnnsenent tt wwawasss ooobtainini eed fffrororom iiinnddivivvididduauallss ttthrhrououugghh tthhehe CCenenntetersrsr ffforrr DDisii eeeasem
CoCoontntntrororoll anandd d PrPPrevevvennntitioonon. .
Laboratoryy MMMetete hohohodsdsds
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
5
fasting individuals.19
Variable Definitions
We classified individuals as fasting if they had fasted for at least 8 hours and stratified
individuals based on fasting status at the time of phlebotomy. The Adult Treatment Panel III
(ATP-III) guidelines define fasting time as 9-12 hours in the US, with new guidelines from the
ACC/AHA taskforce recommending obtaining fasting lipids but not specifying the duration of
fast.1, 20 We used 8 hours to define fasting in keeping with recent studies examining lipids4, 10, 21
and to reflect a more conservative fasting definition. Hypertension was defined as systolic blood
pressure of 140 mmHg, or diastolic blood pressure 90 mmHg as per the 2014 evidence based
hypertension guidelines.22 We defined diabetes as serum glycosylated hemoglobin of 6.5% as
per the World Health Organization’s updated definition of diabetes23 or self-reported history of
diabetes. We used enlarged waist circumference (defined as >88 cm for women and >102 cm for
men) as a proxy for obesity, as waist circumference has been shown to be more highly correlated
with mortality and reflective of central adiposity than BMI.24, 25
Outcome Measures
The primary outcome analyzed was mortality from all-causes and the secondary outcome was
CV mortality. Data on mortality was obtained using death records from the NDI cross-matched
to NHANES-III using probabilistic record matching. International Classification of Diseases,
Ninth Revision (ICD-9) and ICD-10 codes were recoded as underlying classification of death
(UCOD) within the NHANES III-NDI. Deaths from CV related diseases included deaths from
ischemic heart disease (I20-I25), heart failure (I50), essential hypertensive heart disease (I11-
I13), cerebrovascular disease (I60-I69) and atherosclerosis (I70-I71).
hypertension guidelines.22 We defined diabetes as serum glycosylated hemoglobibiin ofoo 6.6.6 5%5%5% aas
per the World Health Organization’s updated definition of diabetes23 or self-reported history of
diiababbeteteteeses. WWWeee ussededed eenlarged waist circumferenceee ((dddefined as >888 cccm fofoforr r wow men and >102 cm fo
mmenn)n) as a proxxy y fooor obobobesessititity,y, aasss wawaaisisst t ccirrccummmfeeerenncecee haasas bbeeeeenn shshowowown n totoo bbee momomorere hhhigigighlhlhly y cooorrrreele aaated
wiwiiththth mmmorortataalililitytyty aandndnd rrefefflelectctc ivive e ofoff cccenenenttrtralall aaadididipopoposisiitytyt ttthahaan BMBMMI.II.fff 244,, 2525
Outcome MeMeeasasasururreseses
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
6
Statistical Analysis
All analyses were performed using SAS software version 9.3 (SAS Institute Inc). We adjusted
for the complex, stratified study sampling design using survey weights for examination and
interview portions of survey as per the CDC recommendations.26 Sensitivity analyses were
performed without using survey weights.
Propensity Score Matching
We used propensity score matching to assemble a cohort of paired participants based on fasting
status with similar baseline characteristics. Propensity score was calculated using a non-
parsimonious multivariable logistic regression model with fasting status (dichotomized as yes or
no) as the dependent variable. CV risk factors were entered into the model as covariables to
control for possible confounders (including race, smoking history, prior CVD, cholesterol
medication use, diabetes, elevated TC, low HDL-C, hypertension, enlarged waist circumference
and low socioeconomic status). Matching was performed using SAS 9.3 and SAS macro
(GMATCH) with greedy matching in a 1 to 1 ratio without replacement, with caliper width of
0.2 times the standard deviation of the logit of the propensity scores. The discriminatory power
of the fasting and non-fasting LDL-C model was evaluated using the area under the receiver-
operator curve (ROC) using the Hosmer-Lemeshow C-statistic. Fasting and non-fasting ROC
curves were compared using bootstrapping methods to evaluate for a statistically significant
difference. Absolute standardized differences were calculated between the fasting and non-
fasting cohort before and after propensity score matching.
We generated Kaplan-Meier curves to assess survival functions in both fasting and non-
fasting cohorts. Primary analysis was performed on the matched cohort. The prognostic values of
fasting vs. non-fasting LDL-C measurement for primary and secondary outcomes were assessed
no) as the dependent variable. CV risk factors were entered into the model as covovvararriaiablblbleseses ttto o o
control for possible confounders (including race, smoking history, prior CVD, cholesterol
memedididicacacatititiononon uussse, dididiababa etes, elevated TC, low HDLLL-C-C- ,, hypertensionn, , enlalaargrgrgede waist circumference
anndd d lol w socioeoecocoonnnommimic stststatatatususus)).). MaMaMatctchhihinnng wwawas peeerffformememed d uuusinini gg SASASS 99.9 33 ananndd d SASASSS mmamacccroo o
GGGMAMAMATCTCT H)H)H) wwitithhh ggrgreeeeedydyy mmatatatchchhinining g g iinin aaa 111 ttto o o 1 rararatitiio wiwiiththouououtt t reeeplplp aacacemememenennt,, wwititith h cacac liliipepeper wwiwidtdtd h h ofoff
0.2 times the e stststanana dadadardrdrd dddeveviaiaiatititiononn ooof f thththe lolologigig t t t ofofof ththt eee prprpropopenenensisisitytyty ssscococoreees.s.s TTThehehe dddisisi crcrrimimminininatatatoro y power
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
7
using ROC curves. Sensitivity analysis to assess whether the prognostic significance of fasting
vs. non-fasting LDL-C varies by length of fast was performed on the unmatched cohort using
different cut-points to define fasting status (<4 vs. 4 hours, <8 vs. 8 hours, <12 vs. 12 hours).
We stratified by presence of diabetes to determine whether diabetic status influenced prognostic
significance of fasting in unmatched models. Further sensitivity analyses were performed
including patients with triglycerides 400 mg/dL. We conducted sensitivity analyses at different
follow up time cut-points (5, 10, and 15 years) to ensure that the significance of fasting status did
not vary by follow-up length. Analyses were also performed to evaluate the influence of fasting
vs. non-fasting TC and triglycerides levels on all-cause and CV mortality.
Cox proportional hazard models were used to evaluate the association of LDL-C levels
with outcomes after adjustment for potential confounders. Individuals were stratified by tertiles
of LDL-C levels (<100 mg/dL [referent], 100-130 mg/dL, and 130 mg/dL) with the lowest
tertile used as the reference group. Secondary analyses were performed using clinical cut points
(LDL-C levels <130 mg/dL, 130 mg/dL-160 mg/dL, and 160 mg/dL). Interaction between
fasting status and LDL-C was tested in both primary and secondary outcome models using
interaction terms for fasting state and LDL-C tertiles. Two tailed p-values of 0.05 or less were
considered statistically significant.
Results
Our initial dataset included 20,024 adults. As shown by Figure 1, we excluded those in whom
LDL-C calculation was not possible (n=699), those with triglycerides 400 mg/dL (n=440) (see
sensitivity analyses below), and those in whom fasting time was missing (n=2,723), or final
mortality status was missing (n=1). Thus, our final dataset included 16,161 individuals
Cox proportional hazard models were used to evaluate the association of f LDLDLDLL--CCC lelelevevevelsls
with outcomes after adjustment for potential confounders. Individuals were stratified by tertiles ff
off LLLDLDLDL CC-C lllevevvelss (<(<(<101 0 mg/dL [referent], 100-1131300 mg/dL, and 131 0 0 mgmgmg/d/ L) with the lowest
eerttiilile used as ththhee rrrefefef rerencncnceee grgrgroououp.p.p. SSSececcoonndarryry anaaalyyysesss wwwerere peperrffoorrmemed dd uususininng gg clclininnicicalalal cccututt ppooio nnnts
LLLDLDLDL-C-C-C llevevvelelelss <<<133030 mmmg/g/g/dLdL, 131313000 mgmmg//d/dL-L-L-1616160 00 mgmgmg/dddL,L, andndnd 1616160 0 mgmgmg/d/d/ LLL).. IIntntntererracactititioonon bbbetttweweeenen
fasting statuss aaandndn LLLDLDLDL-C-C wwwasasas ttesesestetet d d d innn bbbotototh h h prprp imimimararary y y ananndd d sesesecococondndndaraa y y y ouououtctctcomomomee momomodededelslsls uusing
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
8
representing 172,332,619 adults in the US population.
During a mean follow-up of 14.0 (+/- 0.22) years, there were a total of 3,788 deaths
(23.4%) and 1,454 (9.0%) CV deaths. Among the 16,161 individuals 10,023 (62.0%) participants
were fasting, and 6,138 (38.0%) individuals were non-fasting at the time of phlebotomy. Prior to
propensity score matching there were significant differences in the baseline variables between
the two groups (Table 1). Propensity score matching matched 4,299 (42.9% of fasting; 70.0% of
non-fasting) individuals with similar propensity scores. Post matching, there were no significant
differences between the baseline characteristics of the two groups and the absolute standardized
differences were <10% for all matched variables indicating an adequate match.27
All-Cause Mortality
In the unmatched cohort, there was an increased risk of all-cause mortality with increasing LDL-
C tertile [HRs 1 (referent), 1.57 (95% CI 1.34-1.83) (2nd tertile), 2.00 (95% CI 1.70-2.33) (3rd
tertile), respectively]. Test for interaction between fasting status and all-cause mortality was not
significant (Pinteraction = 0.64) indicating lack of association between fasting status and LDL-C
with all-cause mortality (eTable 1). Furthermore, the C-statistics for fasting vs. non-fasting
groups for predicting all-cause mortality were similar [0.58 (95% CI 0.57-0.60) vs. 0.58 (95% CI
0.56-0.59); P =0.55] (eFigure 1) suggesting similar prognostic value of fasting and non-fasting
LDL-C levels. Analyses including individuals with triglycerides of 400 mg/dL did not show a
significant difference between fasting vs. non-fasting C-statistics [0.58 (95% CI 0.57-0.60) vs.
0.57 (95% CI 0.55-0.59); P=0.34] (eFigure 2). Results were largely similar based on diabetic
status: fasting vs. non-fasting C-statistics in non-diabetics were not significantly different [0.59
(95% CI 0.57-0.60) vs. 0.59 (95% CI 0.57-0.61); P=0.79] nor were C-statistics in diabetics
[0.51(95% CI 0.46-0.56) vs. 0.51(95% CI 0.46-0.56); P=0.98] (eFigures 3 and 4).
All-Cause Mortality
n the unmatched cohort, there was an increased risk of all-cause mortality with increasing LDL-
C C teteertrtrtilililee [H[H[HRsRsR 1 (((rererefef rent), 1.57 (95% CI 1.34-1.8.8.83)) (2nd tertile), 222.0. 0 (9(9(955%5% CI 1.70-2.33) (3rd
eerttiilile), respecctitiivvvellly].]. TTeesesttt fofofor r r ininteteeraraactctioionn beeetwwweennn ffastttininng g ssstaatatusus anndnd aallll -ccauauusesee mmororortataaliliittyty wwwasass nnnot
iigngngnififficicicanantt (P(P(Pintintereraractitiionono === 000.6.64)4) iindndndicicicatatatiningg g lalalackckck oof ff asassososocic aaatiiononon bbbetetetweweweenenn fffasasstinngng ssstatatatutut s s ananand LDLDLDL-L-CCC
with all-causesee mmmororrtatatalilitytyty ((eTeTTababablelee 111).).. Fuuurtrtrthehehermrmrmororo e,e,e ttheheh CCC-s-sstatatatitit stststicicics s fofofor r r fafafastststinining g vsvsvs.. nononon-n-n faf sting
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
9
Sensitivity analysis using different cut-point definitions for fasting of <4 hours vs. 4
hours [0.58 (95% CI 0.57-0.59) vs. 0.60 (95% CI 0.56-0.64); P=0.37] or for <12 hours vs. 12
hours [C-statistics 0.57 (95% CI 0.56-0.59) vs. 0.59 (95% CI 0.57-0.60); P=0.37] (eFigures 5
and 6) showed largely concordant results with using an 8 hour fasting cut-point definition, and
did not show significant difference between fasting and non-fasting groups. Sensitivity analysis
using different follow up times did not show significant different between fasting and non-
fasting groups (data not shown).
Within the propensity score matched cohort, there was increased risk of all-cause
mortality by increasing LDL-C tertile [HRs 1 (referent), 1.61 (95% CI 1.25-2.08) (2nd tertile),
2.10 (95% CI 1.70-2.61) (3rd tertile), respectively]. There was no difference between fasting vs.
non-fasting LDL-C and all-cause mortality within each tertile of LDL-C (Figure 2). Test for
interaction between fasting status and all-cause mortality was not significant (Pinteraction = 0.11)
indicating lack of association between fasting status and LDL-C with all-cause mortality (Table
2). Similarly, the C-statistics for the fasting and non-fasting groups for predicting all-cause
mortality were similar [C-statistics 0.59 (95% CI 0.56-0.61) vs. 0.58 (95% CI 0.56-0.60); P=
0.73] (Figure 3).
In the unmatched cohort, C-statistics of triglyceride levels in fasting vs. non-fasting
groups for predicting all-cause mortality were not significantly different [(C-statistics 0.60 (95%
CI 0.59-0.62) vs. 0.61 (95% CI 0.59-0.62); P=0.96, respectively] (eFigure 7). Similarly, C-
statistics of TC level in fasting and non-fasting groups for predicting all-cause mortality were not
significantly different [(C-statistics 0.60 (95% CI 0.59-0.62) vs. 0.59 (95% CI 0.57-0.61);
P=0.31] (eFigure 8).
2.10 (95% CI 1.70-2.61) (3rd tertile), respectively]. There was no difference betwtwweeeen fafafastststinining g g vsv .
non-fasting LDL-C and all-cause mortality within each tertile of LDL-C (Figure 2). Test for
nnteteerararactctctiioionnn bebebetwweeeeeennn fasting status and all-cause momom rrtality was not t sis gnnififificicicant (Pinteraction = 0.11)
nndiiicac ting lack k ofoo aassssooociaiaiatititiononon bbebetwtwweeeeenn fafaastinnngg statttuuss andndnd LLDLDLDL--CC wwiwiththh aaalll-c-cauauaussese mmmororo tatat lililityyy (((TaTaTabbble
2)2).. SiSiSimimim lalarlrlrly,yy, tthhee CCC-s-sttattitiststiccs ss fofoforrr thththee fafaaststtinining gg anannd d d nnononn-ffafastststininingg g grgrgrououupspsps ffoorr pprprededdicicctitit ngngng aaalll-c-ccauauusese
mortality weereree sssimimmilillararr [[[C-C-ssstatatatitt ststticicicss 0.00 59599 (((959595% % % CICICI 000.5.556-6 0.0.0.616161) ) ) vsvsvs.. 0.00 585858 (((959595% % % CICIC 000.5.5.56-6-6 0.0.0 60); P=
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
10
Cardiovascular Mortality
Outcomes for CV mortality prior to propensity score matching similarly demonstrated increased
risk of CV mortality by increasing LDL-C tertile [HRs 1 (referent), 1.82 (95% CI 1.38-2.39) (2nd
tertile), 2.94 (95% CI 2.20-3.93) (3rd tertile)]. Test for interaction between fasting status and all-
cause mortality was not significant (Pinteraction = 0.11) indicating lack of association between
fasting status and LDL-C with CV mortality (eTable 1). Fasting vs. non-fasting C-statistics
were also similar [0.62 (95% CI 0.59-0.64) vs. 0.62 (95% CI 0.60-0.64); P=0.80] (eFigure 9)
suggesting similar prognostic value of fasting and non-fasting LDL-C levels on CV mortality.
Fasting vs. non-fasting C-statistics in non-diabetics were similar [0.62 (95% CI 0.60-0.65) vs.
0.64 (95% CI 0.61-0.67); P=0.42] as well as in diabetics [0.55 (95% CI 0.49-0.61) vs. 0.53 (95%
CI 0.47-0.60); P=0.67] (eFigures 10 and 11). Sensitivity analysis including individuals with
triglycerides of 400 showed largely concordant results with similar prognostic value of fasting
and non-fasting LDL-C levels [0.62 (95% CI 0.60-0.64) vs. 0.61 (95% CI 0.58-0.63); P=0.51]
(eFigure 12).
Sensitivity analysis using different cut-points for fasting of <4 hours vs. 4 hours [0.62
(95% CI 0.60-0.64) vs. 0.65 (95% CI 0.59-0.70); P=0.34] (eFigure 13), or for <12 hours vs. 12
hours [0.61 (95% CI 0.59-0.63) vs. 0.63 (95% CI 0.61-0.66); P=0.27] (eFigure 14) showed
largely concordant results with an 8 hour fasting cut-point definition, showing similar prognostic
value of fasting vs. non-fasting LDL-C. Sensitivity analysis using different follow up time cut-
points did not show significant difference between fasting and non-fasting groups (data not
shown).
In the propensity score matched cohort, there was increased risk of CV mortality by
increasing LDL-C tertile [HR 1 (referent), 1.68 (95% CI 1.13-2.51) (2nd tertile), 3.04 (95% CI
0.64 (95% CI 0.61-0.67); P=0.42] as well as in diabetics [0.55 (95% CI 0.49-0.66111) vvvs. 000.5.55333 (9(9(955%
CI 0.47-0.60); P=0.67] (eFigures 10 and 11). Sensitivity analysis including individuals with
rrigigglylylycecece iriridededes s oof 440400 0 showed largely concordannt t t reressults with simmilililar pprororogggnostic value of fasting
anndd d non n-fastinng g LDLDLDL-L-CC lelelevevelslsls [[0.0.626262 ((99595%% CCCI 00.6000-00.6444))) vsvs. 00.0.61611 ((9595%%% CCCI 0.0.0.58558-0-0.6.6.63)3)3 ;; P=P=P=000.5515 ]]]
eFeFFigigigururu ee 12122).)).
Sensititivivivititi y y y anananalala ysysy isss uuusiss ngngng ddififffef rererentntnt cccuuut-t-t popopoininntstss ffororr fffasasa titit ngngng ooof f <4<4<4 hhhououoursrsrs vvvs.s.s. 4 4 4 hohohours [0.62
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
11
2.00-4.62) (3rd tertile); respectively]. Test for interaction between fasting status and CV
mortality remained non-significant (Pinteraction = 0.34) (Table 2) indicating lack of association
between fasting status and LDL-C with CV mortality. Similarly, the C-statistics for the fasting
and non-fasting groups for predicting CV mortality were similar [0.62 (95% CI 0.60-0.66) vs.
0.62 (95% CI 0.60-0.66); P=0.96] (Figure 4) suggesting similar prognostic value of fasting and
non-fasting LDL-C.
In the unmatched cohort, C-statistics of triglyceride levels in fasting vs. non-fasting
groups predicting CV mortality were not significantly different [(C-statistics 0.62 (95% CI 0.60-
0.64) vs. 0.61 (95% CI 0.59-0.64); P=0.81, respectively] (eFigure 15). The C-statistics of TC
levels for CV mortality in fasting and non-fasting groups were similarly not significantly
different [C-statistics 0.64 (95% CI 0.62-0.66) vs. 0.63 (95% CI 0.60-0.65); P=0.49] (eFigure
16).
Sensitivity analyses without using survey weights yielded largely similar results for both
primary and secondary outcomes (data not shown).
Discussion
Every year millions of blood samples are drawn across the world for the measurement of lipid
panels, in particular LDL-C, with most national and international guidelines recommending a
fasting panel for such measurement. The results of this nationally representative cohort study
with 16,161 individuals followed for 14.0 years representing >172 million adults in the US
population show similar prognostic value of non-fasting LDL-C levels as compared to fasting
LDL-C levels for prediction of both all-cause mortality as well as CV mortality, thereby
questioning this traditional practice.
evels for CV mortality in fasting and non-fasting groups were similarly not signniifificiccannntltltly yy
different [C-statistics 0.64 (95% CI 0.62-0.66) vs. 0.63 (95% CI 0.60-0.65); P=0.49] (eFigure
166).).).
Sensitivivititityy annaala yysyseseses wwwiitithohoututut uussinnng suuurvvvey weweighghghttsts yyyieeldldeded llarargegeelyyy ssimimmiililarar rrresssulululttts ffforor bbboooth
prprimimimarararyy y ananndd d sesecocoonnddararyy oououtct oomomeseses (((dadad tataa nnnototot ssshooownwnwn).
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
12
Fasting Lipid Panel
The origin of the need for a fasting lipid panel is not entirely clear. It is known that certain lipid
parameters, especially triglycerides may be sensitive to fasting status and to the content of the
last meal (and in particular high fat loads). As such, fasting panels have been recommended to
provide accurate lipid measurements. However there are a number of drawbacks with this
approach including the need to reschedule a visit for a separate blood draw if patient is not
fasting thereby decreasing compliance and delaying treatment. Moreover, as individuals are in a
non-fasting state for the majority of time during the day, obtaining a fasting lipid panel may not
accurately reflect post-prandial abnormalities in lipid metabolism and thus obtaining a non-
fasting lipid panel may reflect a more relevant physiological state.7, 8 Obtaining a non-fasting
blood sample may also offer the opportunity to assess non-fasting blood glucose which may add
accuracy in identifying glucose intolerance.28, 29
Recently, several studies have questioned the need for fasting lipid profile, mostly
involving the use of non-fasting triglycerides in CV risk assessment. Although the role of
triglycerides as an independent CV risk factor is less clear than LDL-C, studies have shown that
postprandial triglycerides are similar or possibly even superior to fasting triglycerides in CV risk
prediction.9, 10, 30-32 Recent recommendations suggest potentially moving towards non-fasting
triglycerides for risk assessment, however further research is needed before definitive
recommendations can be made.33, 34
Fewer studies have addressed the use of non-fasting LDL-C in risk prediction. Numerous
animal, population based, and clinical studies have shown that LDL-C is associated with
increased CV mortality,35-37 with genetic studies also showing a causative mortality linkage.38-40
These studies have traditionally used fasting LDL-C as convention and thus recommendations
fasting lipid panel may reflect a more relevant physiological state.7, 8 Obtaining aaa nonoo -f-f-fassastititingngng
blood sample may also offer the opportunity to assess non-fasting blood glucose which may add
acccucuuraraaccycy iiinn n idididenntititiffyfying glucose intolerance.28, 299
Recentntlyly,, ssseveveeraral ll stststudududieieies s hhahavveve qquuuestiioi nnned thhhe nneneeeded fffoorr ffaaasttiingng lllipipidid ppproroofifilelee,, mmomossts lylyly
nnvovovolvlvlvinini gg g thththee e ususeee oofof nnononn-f-fasa tititingngng trtrtrigigiglylycceceririridededes ininin CCCVV V riiiskskk aaasssssesesssmsmsmenennt.t.t. AAAltthhohougugughh h ththe ee rrroleee ooof f
riglycerides s asasas aan n n ininndededepepep ndndndenenent CVCVCV rrrisi k k k fafafactcttororor iiis s lelelesssss ccleleeararar ttthahahan n n LDLDLDL-L-L-C,C,C, ssstututudidid esese hhhavavave e e shs own thatt
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
13
made by various agencies such as the ATP-III have generally been for obtaining fasting lipids.
However, multiple trials, including the Heart Protection Study and Anglo-Scandinavian Cardiac
Outcomes Trial included individuals who were not fasting during the time of phlebotomy when
analyzing effects of lipid lowering agents, suggesting that some of the data supporting lipid
lowering therapy actually springs from studies involving non-fasting individuals. 41, 42
Prior studies examining CV events have demonstrated increased CV risk by LDL-C level
for individuals in a non-fasting state,21, 43-45 but none have examined long-term mortality
outcomes in a representative sample. A recent population based study by Sidhu et al in 2012
showed that in a population based sample, lipid levels by subclass varied little with respect to
fasting time, and by less than 10% for LDL-C.4 Other studies have also shown little variation
with post-prandial LDL-C levels when compared with fasting levels.44, 46 These studies suggest
that the variation between fasting and non-fasting LDL-C levels, if any, is small. Our study is the
first to show that in a population-based sample, the association between LDL-C with CV and all-
cause mortality does not differ by fasting status. Our analyses also suggest that obtaining fasting
TC and triglyceride levels do not have improved prognostic significance over that of non-fasting
levels.
This provides further evidence that it may be unnecessary to use fasting lipid levels to
risk stratify patients. In our primary analyses we excluded patients with a triglyceride level 400
mg/dl, or roughly ~2% of the total population. However, the results were largely concordant in a
sensitivity analysis after including the above patients. Thus, our results are broadly applicable to
all patients undergoing blood draw to assess lipid panel and is applicable to LDL-C measurement
as well as triglyceride and TC measurement.
fasting time, and by less than 10% for LDL-C.4 Other studies have also shown liliittttlelee vvaraariaiaiatititionono
with post-prandial LDL-C levels when compared with fasting levels.44, 46 These studies suggest
hhatatt ttthehehe vvararariaiaiatttionnn bbbete ween fasting and non-fastingngng LLLDL-C levels, iiif f anny,y,y iiis s small. Our study is the
fifiirstt t to show ththatatt innn a a popooppupulalalatititioonon-b-bbaasasededd ssamplplp ee, thhhee assososociciaatatiooon n bbebetwtweeeenn LDLDDL-L-L-CC wwiwiththth CCCV V V anannd d aall-
caausususe ee momom rtrttalalaliitity y dododoeses nnnotott ddififfefefer bybyby fffaasastitiingngng ssstatatatuuus.s.s uOuOur anananalalalyysyseseses aallsl ooo susus gggggesese t t ththt atata ooobtbtbtaainniningngg ffaaastitiingn
TC and triglycycycerererididde e e leleevevevels dddooo nooot t t hahahavevv iiimpmpmprororoveveved d prprprogogognonoostststicicic ssigigigninin fiicacac ncncnce e e ovovoveer r thththatatat ooof f f non n-fastinggg
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
14
2013 ACC/AHA Guidelines and LDL-C Measurement
The recently published 2013 ACC/AHA guidelines recommend obtaining fasting lipids, however
the guidelines do not specify the length of time for fasting, nor cite data to support the need for
fasting LDL-C. The guidelines move away from recommending lowering LDL-C to specific
targets but recommend moderate to high intensity statin for patients with atherosclerotic CV
disease, an intensity of statins that would reduce baseline LDL-C by approximately 40-50%
which can easily be assessed with a non-fasting sample.
This has important implications in clinical practice. Requiring patients to fast causes
patients increased stress, potential hypoglycemia in patients with diabetes, increased
transportation costs, and potentially missed days of work. In addition, the inconvenience of
fasting may also delay treatment or diagnosis of hyperlipidemia if patients are unable to fast
before clinic visits. Enabling patients to obtain non-fasting lipid profiles would improve patient
satisfaction, and potentially avoid delays in detection and treatment of hyperlipidemia while at
the same time providing similar prognostic value as that of a non-fasting LDL-C value.
Limitations
The design of this study using data from an existing database limits our ability to prove that
fasting and non-fasting lipids have the same prognostic value. In addition, fasting and non-
fasting LDL-C were not collected on the same individuals. Moreover, in non-fasting patients
data was not available on the composition of patient meals.
Conclusions
In conclusion, the results of this study of 16,161 individuals followed for 14.0 years and
representative of the US population fail to show a superior prognostic value of fasting LDL-C
ransportation costs, and potentially missed days of work. In addition, the inconvnvvennnieiencncn e ee ofofof
fasting may also delay treatment or diagnosis of hyperlipidemia if patients are unable to fast f
beefofoforeree cclililinininicc vvisisiitststs. Enabling patients to obtain nonon nnn-fasting lipid prpp offililleeess s would improve patient
aatiissfsfaction, anand dd pootetennntiaiaalllllly y avavavoioid dd dededelalaayyss in deeetectttiooon aaanndnd ttrrreatatmmmenntnt ooff f hyyypepeerlrlrliipipididdemememiaia wwhihihileele aaatt
hhe e e sasaamemem ttimimimee prprrovvvididiinng g g sisimimiilalar rr prprprogoognonoostststicicic vvvalalaluuuee asss thhahattt ofofof a nnnononn-fffasasastitingngng LLDLDLDL-C-C vvvaalalueee.
Limitations
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
15
levels when compared with non-fasting LDL-C levels both for the prediction of all-cause
mortality as well as CV mortality. Our study suggests that a non-fasting LDL-C measurement
offers a more convenient method of phlebotomy while preserving the prognostic value of the
test. National and international guideline societies should re-consider the need for fasting LDL-
C. Similar results were seen for triglycerides and TC thus questioning the value of obtaining
fasting lipid profile.
Acknowledgments: Data analysis and statistical support were provided by New York University
School of Medicine Cardiovascular Outcomes Group. Author Contributions: Authors YG, BD,
and SB had full access to all of the data in the study and take responsibility for the integrity of
the data and the accuracy of the data analysis. Study concept and design: BD and SB.
Analysis and interpretation of data: YG, JX, BD and SB. Drafting of the manuscript: BD and
SB. Critical revision of the manuscript for important intellectual content: BD, SM, HW, DM and
SB. Statistical analysis: YG, JX, BD and SB
Conflict of Interest Disclosures: Dr. Mora reports research grants from Atherotech Diagnostics,
AstraZeneca and is on the advisory board for Quest Diagnostics, Cerenis Therapeutics,
Genzyme, Lilly. Dr. Bangalore is on the advisory board for Pfizer. The remaining authors have
no conflicts to disclose.
References:
1. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002;106:3143-3421. 2. De Backer G, Ambrosioni E, Borch-Johnsen K, Brotons C, Cifkova R, Dallongeville J, Ebrahim S, Faergeman O, Graham I, Mancia G, Manger Cats V, Orth-Gomer K, Perk J, Pyorala K, Rodicio JL, Sans S, Sansoy V, Sechtem U, Silber S, Thomsene T, Wood D. European guidelines on cardiovascular disease prevention in clinical practice. Third Joint Task Force Of European and other societies on cardiovascular disease prevention in clinical practice (constituted by representatives of eight societies and by invited experts). Arch Mal Coeur Vaiss. 2004;97:1019-1030.
y p y gg yy
he data and the accuracy of the data analysis. Study concept and design: BD andndd SSB.BB
Analysis and interpretation of data: YG, JX, BD and SB. Drafting of the manuscript: BD and
SB. CrC itical revision of the manuscript for impoortant intellectual content: BD, SM, HW, DM and
SBSBB. SStStata isstiticacacal ananalalysysiss:: YGYG, JXJX, BDD aandd SB
CoCoonfnfnflil ctc of InInInteerererestt DDDisssclossururreses: DDDr.r. MMMororaa reeepoporrrtss rereeseseeararcchh ggrarannttss frorom mm AtAttheheerorotetechchch DDDiaaagngnooosttiics
AstrtraZaZenececa anand d iss oon n thhe e addvivisosoryr boardrd ffoor r QuQuest DiDiagagnonoststicics, CCererenenis TTheherarapep uticics,s
Genzymy e, LLililillylyly. DrDrDr. BaBaBanngngalalalororo e isisi oonnn thhheee adadadvivivisososoryryry bbboaoaarddd ffforor PPPfififizezer. TTThehehe rrremememaiainininingngg aaautututhors haveyy
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
16
3. Genest J, McPherson R, Frohlich J, Anderson T, Campbell N, Carpentier A, Couture P, Dufour R, Fodor G, Francis GA. 2009 canadian cardiovascular society/canadian guidelines for the diagnosis and treatment of dyslipidemia and prevention of cardiovascular disease in the adult–2009 recommendations. Can J Cardiol. 2009;25:567-579.
4. Sidhu D, Naugler C. Fasting time and lipid levels in a community-based populationa cross-sectional studyfasting time and lipid levels. Arch Intern Med. 2012;172:1707-1710. 5. Langsted A, Nordestgaard BG. Nonfasting lipids, lipoproteins, and apolipoproteins in individuals with and without diabetes: 58 434 individuals from the Copenhagen General Population Study. Clin Chem. 2011;57:482-489. 6. Karpe F. Postprandial lipoprotein metabolism and atherosclerosis. J Intern Med. 1999;246:341-355. 7. Zilversmit DB. Atherogenesis: A postprandial phenomenon. Circulation. 1979;60:473-485. 8. Kolovou GD, Anagnostopoulou KK, Daskalopoulou SS, Mikhailidis DP, Cokkinos DV. Clinical relevance of postprandial lipaemia. Curr Med Chem. 2005;12:1931-1945. 9. Nordestgaard BG, Benn M, Schnohr P, Tybjærg-Hansen A. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. JAMA. 2007;298:299-308. 10. Bansal S, Buring JE, Rifai N, Mora S, Sacks FM, Ridker PM. Fasting compared with nonfasting triglycerides and risk of cardiovascular events in women. JAMA. 2007;298:309-316. 11. Ridker PM. Fasting versus nonfasting triglycerides and the prediction of cardiovascular risk: Do we need to revisit the oral triglyceride tolerance test? Clin Chem. 2008;54:11-13. 12. Langsted A, Freiberg J, Tybjærg Hansen A, Schnohr P, Jensen GB, Nordestgaard B. Nonfasting cholesterol and triglycerides and association with risk of myocardial infarction and total mortality: The Copenhagen City Heart Study with 31 years of follow-up. J Intern Med. 2011;270:65-75. 13. Third National Health and Nutrition Examination Survey (NHANES III). Nhanes III household adult data file documentation - ages 17+. 1996. 14. Myers G, Cooper G, Winn C, Smith S. The Centers for Disease Control-National Heart, Lung and Blood Institute Lipid Standardization Program. An approach to accurate and precise lipid measurements. Clin Lab Med. 1989;9:105. 15. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499-502.
8. Kolovou GD, Anagnostopoulou KK, Daskalopoulou SS, Mikhailidis DP, Cokkkkikinonn ss DVDVDV. .Clinical relevance of postprandial lipaemia. Curr Med Chem. 2005;12:1931-19445.5
9. Nordestgaaard BGG, Benn M, Schnohr P, Tybjæærgg-Hansen A. Nonfasting triglycerides and risk off mmmyoyoyocacardrdrdiiaial l innfafafarcrction, ischemic heart disease,e, aannnd death in meen nn annd d d wwowomen. rr JAMA. 2020007077;2298:22999999-33080808.
10100. BBaB nsal SS,, Buuuririnng JJJEE,, RRififfaiii N, MoMoM rraa SS, SSacccks FMFMFM,, RiRiRidkdkdkeerer PMMM. FFasttiinngg cccommmpap reeedd wwiithhh nononfnfnfasasastitit ngngg tttriririglglyyyceereridideses aandn rrrisiskkk ofofof ccararrdidiiovovovaaascucuculalaarr eeveveenentststs iiinnn wowowomememen.n.n. JAJAJAMMAMA. 2020007077;2;2;29898:::300909-3-331666.
11. Ridker PPM.M.M FFasasastitit ngngn vvererrsusususs nonononfnfnfasaa titiingngng ttriririglglglycycycerereriideded s ananand d d thththe e e prprp edddicicictitit ononon oooff cccararardididiovovovasaa cular risk::ffDoDo wwee neneeded ttoo rerevivisisitt ththee ororalal ttririglglycycererididee totolelerarancncee tetestst?? ClClinin CChehemm 20200808;5;54:4:1111 1-133
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
17
16. Miller WG, Myers GL, Sakurabayashi I, Bachmann LM, Caudill SP, Dziekonski A, Edwards S, Kimberly MM, Korzun WJ, Leary ET. Seven direct methods for measuring HDL and LDL cholesterol compared with ultracentrifugation reference measurement procedures. Clin Chem. 2010;56:977-986. 17. Schectman G, Patsches M, Sasse E. Variability in cholesterol measurements: Comparison of calculated and direct LDL cholesterol determinations. Clin Chem. 1996;42:732-737. 18. Nauck M, Warnick GR, Rifai N. Methods for measurement of LDL-cholesterol: A critical assessment of direct measurement by homogeneous assays versus calculation. Clin Chem. 2002;48:236-254. 19. Mora S, Rifai N, Buring JE, Ridker PM. Comparison of LDL cholesterol concentrations by Friedewald calculation and direct measurement in relation to cardiovascular events in 27 331 women. Clin Chem. 2009;55:888-894. 20. Stone NJ, Robinson J, Lichtenstein AH, Bairey Merz CN, Lloyd-Jones DM, Blum CB, McBride P, Eckel RH, Schwartz JS, Goldberg AC, Shero ST, Gordon D, Smith SC, Jr., Levy D, Watson K, Wilson PW. 2013 ACC/AHA guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63:2889-2934 21. Mora S, Rifai N, Buring JE, Ridker PM. Fasting compared with nonfasting lipids and apolipoproteins for predicting incident cardiovascular events. Circulation. 2008;118:993-1001. 22. James PA OS, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J, Lackland DT, LeFevre ML, MacKenzie TD, Ogedegbe O, Smith SC Jr, Svetkey LP, Taler SJ, Townsend RR, Wright JT Jr, Narva AS, Ortiz E. 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;311:507-20. 23. World Health Organization. Use of glycated haemoglobin (Hba1c) in the diagnosis of diabetes mellitus. Abbreviated Report of a WHO Consultation. 2011. 24. Lean M, Han T, Morrison C. Waist circumference as a measure for indicating need for weight management. BMJ. 1995;311:158-161. 25. Klein S, Allison DB, Heymsfield SB, Kelley DE, Leibel RL, Nonas C, Kahn R. Waist circumference and cardiometabolic risk. Diabetes Care. 2007;30:1647-1652. 26. Centers for Disease Control and Prevention. NHANES tutorial website. http://www.cdc.gov/NCHS/Tutorials/Nhanes/index.htm. Accessed December 10, 2014. 27. Cohen J. Statistical power analysis for the behavioral sciences. Toronto: Academic Press, Inc.; 1977.
McBride P, Eckel RH, Schwartz JS, Goldberg AC, Shero ST, Gordon D, Smith SCCC, , , JrJrJ .,.,, LLevevy y y DD, Watson K, Wilson PW. 2013 ACC/AHA guideline on the Treatment of Blood ChChholollese teteterororol l l tototo Reduce Atherosclerotic Cardiovascular Risk in Adults: A Report of the Americaan n CoCoCollllegegegeee ofofof Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;6;63:288989-293934
21211. MMoMora SS, , RRRifafaii i N,N,N BBururu inining gg JEJE, , RiRiR dkdkerer PPM.M. FaFaststiinnggg coompmpmpararared wwwittthh nononfnfnfaastitingngng lllipipipidids s anannd d dappoollipi oproteinnss fofof rr prprededdicicictititingngng iinnciciidededenntt ccardidiiovvvasccucullar evevvenentsts.. CiCiircccululatattioionn. . 20202000808;1;1118188:9:9993933-1-100000 11.1
2222. JaJaamemem ss PAPAPA OOSSS, CCararrteeer r BLBLL,, CuCuCushshshmmamann n WCWCWC, DDeDennnnniiisoson-n-n HiHiHimmmmmmelelfafaf rbrbrb CC,, HHaHandndndleleer r J,J,J, LLLaccklklanana dd d DDTDT,LeFeFevrv e MLML, MMacKcKenenziee TDTD, OgOgededegbebe OO, SmSmitithh SCC JJr,r SSvevetktkeyy LLP,P, Talalerer SSJ,J TTowwnsn ennd d RRRR, , Wright JT Jrr,, NaNaNarvrvva a a ASASAS,, OrOrOrtitiiz zz E.E.E. 2201010 44 evevevidididenenencecece-b-bbasasasedede ggguiuiuidededeliiinenene fororor ttthehehe mmmanana agagagemememenenentt t of high blbloooodd prpresessusurere iinn adadulultsts:: ReRepoportrt frfromom tthehe ppananelel mmemembebersrs aappppoiointnteded ttoo ththee EiEighghthth JJoiointnt NNatatioionanalldd
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
18
28. Bonora E, Muggeo M. Postprandial blood glucose as a risk factor for cardiovascular disease in type ii diabetes: The epidemiological evidence. Diabetologia. 2001;44:2107-2114. 29. Cavalot F, Petrelli A, Traversa M, Bonomo K, Fiora E, Conti M, Anfossi G, Costa G, Trovati M. Postprandial blood glucose is a stronger predictor of cardiovascular events than fasting blood glucose in type 2 diabetes mellitus, particularly in women: Lessons from the San Luigi Gonzaga Diabetes Study. J Clin Endocrinol Metab. 2006;91:813-819. 30. D Kolovou G, P Mikhailidis D, Kovar J, Lairon D, G Nordestgaard B, Chye Ooi T, Perez-Martinez P, Bilianou H, Anagnostopoulou K, Panotopoulos G. Assessment and clinical relevance of non-fasting and postprandial triglycerides: An expert panel statement. Curr Vasc Pharmacol. 2011;9:258-270. 31. McBride PE. Triglycerides and risk for coronary heart disease. JAMA. 2007;298:336-338. 32. Wild SH, Fortmann SP, Marcovina SM. A prospective case-control study of lipoprotein (a) levels and apo (a) size and risk of coronary heart disease in Stanford Five-City Project participants. Arterioscler Thromb Vasc Biol. 1997;17:239-245. 33. Miller M, Stone NJ, Ballantyne C, Bittner V, Criqui MH, Ginsberg HN, Goldberg AC, Howard WJ, Jacobson MS, Kris-Etherton PM. Triglycerides and cardiovascular disease a scientific statement from the American Heart Association. Circulation. 2011;123:2292-2333. 34. Kannel WB, Vasan RS. Triglycerides as vascular risk factors: New epidemiologic insights for current opinion in cardiology. Curr Opin Cardiol. 2009;24:345. 35. Gordon T, Kannel WB, Castelli WP, Dawber TR. Lipoproteins, cardiovascular disease, and death: The Framingham study. Arch Intern Med. 1981;141:1128. 36. Bush TL, Barrett-Connor E, Cowan LD, Criqui MH, Wallace RB, Suchindran C, Tyroler H, Rifkind BM. Cardiovascular mortality and noncontraceptive use of estrogen in women: Results from the Lipid Research Clinics Program Follow-up Study. Circulation. 1987;75:1102-1109. 37. Kannel W, Neaton J, Wentworth D, Thomas H, Stamler J, Hulley S, Kjelsberg M. Overall and coronary heart disease mortality rates in relation to major risk factors in 325,348 men screened for the MRFIT. Multiple Risk Factor Intervention Trial. Am Heart J. 1986;112:825-836. 38. Eichner JE, Dunn ST, Perveen G, Thompson DM, Stewart KE, Stroehla BC. Apolipoprotein E polymorphism and cardiovascular disease: A huge review. Am J Epidemiol. 2002;155:487-495. 39. Marenberg ME, Risch N, Berkman LF, Floderus B, de Faire U. Genetic susceptibility to death from coronary heart disease in a study of twins. N Engl J Med. 1994;330:1041-1046. 40. Umans-Eckenhausen MA, Sijbrands EJ, Kastelein JJ, Defesche JC. Low-density lipoprotein
participants. Arterioscler Thromb Vasc Biol. 1997;17:239-245.
33. Miller M, Stone NJ, Ballantyne C, Bittner V, Criqui MH, Ginsberg HN, Golldbdbbererergg ACACAC,Howard WJ, Jacobson MS, Kris-Etherton PM. Triglycerides and cardiovascular disease a cientific statement from the American Heart Association. Circulation. 2011;123:2292-2333.
34344. KKaKannell WWWB,B,, VVVasasanann RRRS.S.S. TTririglglglycycycereridideses aass vavascs uululaar risisskkk fafafactc ororrsss: NNewew eeepip dedemimimiololo ogogicic iiinsnsn igigighthth ssffoor cuc rrent opininioioi nnn inin ccararardididiololologoogy.y. CCuCurrrr OOOpinnn CCCardddiiool. 22200000999;22424:3:345455.
3555. GoGoGordrdononn TTT, , KKaKannnnnelell WWWB,B CCCasasstetetelllllliii WWWP,P,P, DDDawawawbebeb r r TTRR. LiLiipopopoprprprotototeieiinsnss,, cacaardrddioiovavavascsccululararar ddissseaeaasese, , anannd ddeatath:h: TThehe FFraamim ngnghaham ststududy.y. ArArchc Intterernn MeMedd.. 198181;1;14141:1:112128..
3636 BuBushsh TTLL BBararreretttt C-Cononnonorr EE CCowowanan LLDD CCririququii MMHH WWalallalacece RRBB SSucuchihindndrarann CC TTyryrololerer HH
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
19
receptor gene mutations and cardiovascular risk in a large genetic cascade screening population. Circulation. 2002;106:3031-3036. 41. Sever PS, Dahlöf B, Poulter NR, Wedel H, Beevers G, Caulfield M, Collins R, Kjeldsen SE, Kristinsson A, McInnes GT. Prevention of coronary and stroke events with atorvastatin in hypertensive patients who have average or lower-than-average cholesterol concentrations, in the Anglo-Scandinavian Cardiac Outcomes Trial—Lipid Lowering Arm (ASCOT-LLA): A multicentre randomised controlled trial. The Lancet. 2003;361:1149-1158. 42. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20536 high-risk individuals: A randomised placebo-controlled trial. Lancet. 2002;360:7-22. 43. Di Angelantonio E, Sarwar N, Perry P, Kaptoge S, Ray KK, Thompson A, Wood AM, Lewington S, Sattar N, Packard CJ. Major lipids, apolipoproteins, and risk of vascular disease. JAMA. 2009;302:1993-2000. 44. Langsted A, Freiberg JJ, Nordestgaard BG. Fasting and nonfasting lipid levels influence of normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction. Circulation. 2008;118:2047-2056. 45. Talmud PJ, Hawe E, Miller GJ, Humphries SE. Nonfasting apolipoprotein b and triglyceride levels as a useful predictor of coronary heart disease risk in middle-aged UK men. Arterioscler Thromb Vasc Biol. 2002;22:1918-1923. 46. Van Dieren S, Nöthlings U, Van Der Schouw Y, Spijkerman A, Rutten G, Sluik D, Weikert C, Joost H, Boeing H, Beulens J. Non-fasting lipids and risk of cardiovascular disease in patients with diabetes mellitus. Diabetologia. 2011;54:73-77.
normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk k prprp ededicictitionon. . Circulation. 2008;118:2047-2056.
45. Talmud PJ, Hawe E, Miller GJ, Humphries SE. Nonfasting apolipoprotein b and triglyceride evels as a useful ppredictor of coronary heart disease risk in middle-agged UK men. y Arterioscler
ThThrororombmbmb VVVasasascc Biiololol.. 202 02;22:1918-1923.
4466. VVan Dierenen SSS, NöNöththlililingngngs s s U,U,U VVVananan DDDeerr Scchhooouww YYY, SSSpipipijkjkeeermmamannn AA,A, RRuuuttteen n G,G,G SSluluuikikk DDD, WWeWeikikikeeert CC,C, JJJooo st H, BoBoeiingngg H,, BBBeulleenensss J. NNonono -f-faaastinngng lipiiddsss anannd d d riririsksks oof ccacarrdiovvavascscuuulaarar ddiseeaeasese inn n patttieeentswiwiiththth dddiaiai bebetetetess s memeelllititi usus. DDiDiababeetetololologogogiaiaia.. 22020111111;5;5;54:4::73737 --7-777.
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
20
Table 1. Baseline Characteristics of Fasting and Non-Fasting LDL-C Cohorts, Pre and Post-Propensity Score Matching
Pre-match Post-matchClinical Variables Fasting
(n=10,023)**Non-fasting (n=6,138)** P Value
ASD(%)
Fasting(n=4,299)**
Non-fasting (n=4,299)** P value
ASD(%)
Age, mean (SE), y 42.68 (0.43) 45.43 (0.55) <0.001 6.4 42.7 (0.54) 43.43 (0.57) 0.20 1.8 Women (%) 51.35 54.07 0.02 5.5 51.66 53.13 0.36 2.9 White (%) 75.38 77.17 0.06 4.2 76.00 77.04 0.41 2.5 Enlarged waist circ (%) 34.44 37.54 0.01 6.5 33.03 34.59 0.27 3.3 Smoker (%) 52.23 53.24 0.29 2.0 52.69 52.55 0.93 0.3 DM (%) 4.93 8.39 <0.001 13.9 5.76 6.88 0.08 4.6 HTN (%) 15.98 18.82 <0.001 7.5 17.42 16.47 0.45 2.5 Elevated cholesterol (%) 26.62 28.66 0.07 4.6 25.89 26.78 0.52 2.0 Prior CVD (%) 4.93 6.30 <0.001 6.0 4.55 4.74 0.75 0.9 Cholesterol lowering med (%) 3.08 3.69 0.16 3.4 3.48 3.51 0.96 0.2 Low SES (%) 13.41 12.24 0.22 3.5 12.07 12.57 0.68 1.5 LDL-C, mean (SE), mg/dL* 125.04 (0.77) 123.71 (0.84) 0.07 1.9 118.55 118.33 0.84 0.3 Low HDL-C (%) 35.74 35.09 0.54 1.4 33.12 33.39 0.85 0.6 Abbreviations: ASD = absolute standardized difference; CVD = cardiovascular disease; med = medication; DM = diabetes mellitus; HDL-C = high density lipoprotein cholesterol; HTN = hypertension; circ = circumference; LDL-C = low density lipoprotein C; SES = socioeconomic status *To convert to mmol/L, multiply values by 0.0259 **N reported based on unweighted numbers; P-values based on weighted values
ged waist circ (%) 34.44 37.54 0.01 6.5 33.03 34.599 00.2.27 7 3ker (%) 52.23 53.24 0.29 2.0 52.69 52.2.555555 00.9.9333 0%) 4.93 8.39 <0.001 13.9 5.76 6 8.8888 0.00.080808 444(%) 15.98 18.82 <0.001 7.5 17.42 16.47 0.45 2
ated cholesterol (%) 26.62 28.66 0.07 4.6 25.89 26.78 0.52 2CVDD (%(%)) 4.93 6.30 <00.0. 01 6.0 4.55 4.74 0.75 0
essteteroror l l lololoweririringngn mm dded (((%)%)% 3.08 3.69 0.0 16 3.4 33.48 8 3.51 0.96 0ESESES S (%(%) ) 1313.4411 1212.2.24 4 0.0 22 33.55 1112.2 0707 112.2 5757 00.6. 8 1
C,CC mmmean (SE), mgmg/d/d/dL*L* 11252525.0.004 4 (0(0(0.77 )7)7) 11 3232 7.71 (0(0.8.8. )4)4 0.0 07 11.9.. 11818.5.5555 11111 888.333333 000 8.8.84 44 0HDHDH L-LL C (%) 5535.74 4 5353 .09 0.0 54 111.4.. 333.12 33. 9939 00.85 0vviaiaiationonons: ASD = aabsbsbsolutee tstana da drdr zizede dififfefef erencncn e;e CVDVDD == cc rarardio avascs uluu ar d sisi aeasese;; memem d d d == memm idid cationoo ; DMD = dd aiabebetetess memem llliti us; DHDHDL-C CC == high dd nnensityott ieieinn n hchcholololesestet rorol;l;l; HHHTNTN == hypypy erertetensn oioion;n cic rc == cirirircucucumfmfmfererenencecc ; LDLDLDL-L-L-C === lolol ww edensn ityy y ilipopopoprprproto ieinn n C;C;C SS SESES === ss cococio cece onon momomicic sstatatuutussonverert t toto mmmmomomol/l/l/L,LL mulultititi lplplyy vavalulueses bbbyy 0.0.0 020 599 eepopop rtrteded bbasaseded oonn ununweweigigghthteded nnumumbebersrs;; PP-vvalalueuess babasesed d onon wweieighghg teted d vavalulueses
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
21
Table 2. Cox Proportional Hazards Model of All Cause and Cardiovascular Mortality by LDL-C Level in Fasting and Non-Fasting Cohorts in the Matched Participants
Fasting Non-FastingOutcome LDL-C range
(mg/dL)*Hazard Ratio
(95% CI) P Value LDL-C range (mg/dL) *
Hazard Ratio (95% CI) P Value
All-Cause Mortality LDL-C 1st tertile 99.60 1 (referent) 100.73 1 (referent) LDL-C 2nd tertile 99.60-129.68 1.61 (1.25-2.08) <0.001 100.73-129.22 1.21 (0.92-1.60) 0.17 LDL-C 3rd tertile 129.68-361.40 2.10 (1.70-2.61) <0.001 129.22-438.00 2.23 (1.76-2.83) <0.001 LDL-C x Fasting Status 0.11 CV Mortality LDL-C 1st tertile 99.60 1 (referent) 100.73 1 (referent) LDL-C 2nd tertile 99.60-129.68 1.68 (1.13-2.51) 0.012 100.73-129.22 1.59 (0.97-2.61) 0.063 LDL-C 3rd tertile 129.68-361.40 3.04 (2.00-4.62) <0.001 129.22-438.00 4.00 (2.58-6.19) <0.001 LDL-C x Fasting Status 0.34
Abbreviations: CV = cardiovascular mortality; LDL-C = low density lipoprotein cholesterol *To convert to mmol/L, multiply values by .0259
( ) ( )L-C 3rd tertile 129.68-361.40 2.10 (1.70-2.61) <0.001 129.22-438.00 2.23 (1.76-2 8.883)3)3) <<<00.0.0000 1L-C x Fasting Status 0.11 Mortality
L-C 1st tertile t 99.60 1 (referent) 100.73 1 (referent) L-C 2nd tertile 99.60-129.68 1.68 (1.13-2.51) 0.012 100.73-129.22 1.59 (0.97-2.61) 0.063 L-C 3rd tertile 129.68-361.40 3.04 (2.00-4.62) <<0.001 129.22-438.00 4.4 00 (2.58-6.19) <0.001L-C C xx x FaFaFastststinininggg StStStataa uss 0.0 433 vviaaiatiti nonons:s: CCV = cacardrdr ioi vascscululara mmorrtatalityty; ; LDL L-L C C = low w densn ity y lipop prototein n hcc lolo esterool lonvnn rrert t tot mmol/L,L, mmmultitiplplplyy y vavalulueseses bbby y y 0.0025252599
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.114.010001
22
Figure Legends:
Figure 1. Participant flowchart.
Figure 2. Kaplan-Meier curve for fasting vs. non-fasting LDL-C levels and all-cause mortality.
Figure 3. Prognostic value of fasting vs. non-fasting LDL-C levels on all-cause mortality in the
matched cohort.
Figure 4. Prognostic value of fasting vs. non-fasting LDL-C levels on cardiovascular mortality
in the matched cohort.
Figure 4. Prognostic value of fasting vs. non-fasting LDL-C levels on cardiovassccculalalar momomortrtrtalalaliitity y
n the matched cohort.
at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
Figure 1 at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
Figure 2 at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
Figure 3 at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
Figure 4 at Karolinska Inst on July 14, 2014http://circ.ahajournals.org/Downloaded from
1
SUPPLEMENTAL MATERIAL
Supplemental Table 1 – All Cause and Cardiovascular Mortality by LDL-C Level in Fasting and Non-
Fasting Cohorts Prior to Propensity Score Matching
Outcome
Fasting Non-Fasting
LDL-C
(mg/dL)*
Hazard Ratio
(95% CI) P Value
LDL-C
(mg/dL)*
Hazard Ratio
(95% CI) P Value
All-Cause Mortality
LDL-C 1st tertile
(reference) ≤106.41 1 (referent)
≤104.76 1 (referent).
LDL-C 2nd
tertile 106.41-138.28 1.57 (1.34-1.83) <0.001 104.76-137.11 1.28 (1.05-1.55) 0.015
LDL-C 3rd
tertile 138.28-380.00 2.00 (1.70-2.33) <0.001 137.11-438.00 2.07 (1.67-2.57) <0.001
LDL-C x Fasting
Status 0.11
CV Mortality
LDL-C 1st tertile
(reference) ≤106.41 1 (referent)
≤104.76 1 (referent)
LDL-C 2nd
tertile 106.41-138.28 1.82 (1.38-2.39) <0.001 104.76-137.11 1.62 (1.15-2.28) <0.001
LDL-C 3rd
tertile 138.28-380.00 2.94 (2.20-3.93) <0.001 137.11-438.00 6.18 (2.21-4.76) <0.001
LDL-C x Fasting
Status 0.64
Abbreviations: CV = cardiovascular; LDL-C = low density lipoprotein cholesterol
*To convert to mmol/L, multiply values by 0.0259
2
FIGURE LEGENDS
Figure 1 - Prognostic value of fasting vs. non-fasting LDL-C level on all-cause mortality in the
unmatched cohort
Figure 2 - Sensitivity analysis: Prognostic value of fasting vs. non-fasting LDL-C level
including patients with triglycerides ≥400 mg/dL on all-cause mortality in the
unmatched cohort
Figure 3 - Prognostic value of fasting vs. non-fasting LDL-C level on all-cause mortality in
patients without diabetes in the unmatched cohort
Figure 4 - Prognostic value of fasting vs. non-fasting LDL-C level on all-cause mortality in
diabetic patients in the unmatched cohort
Figure 5 – Sensitivity Analysis: Prognostic value of fasting (<4 hours) vs. non-fasting (≥4 hours)
LDL-C level on all-cause mortality in the unmatched cohort
Figure 6 - Sensitivity Analysis: Prognostic value of fasting (<12 hours) vs. non-fasting (≥12
hours) LDL-C level on all-cause mortality in the unmatched cohort
Figure 7 - Prognostic value of fasting vs. non-fasting triglyceride level on all-cause mortality in
the unmatched cohort
Figure 8 –Prognostic value of fasting vs. non-fasting total cholesterol level on all-cause
mortality in the unmatched cohort
Figure 9 - Prognostic value of fasting vs. non-fasting LDL-C level on cardiovascular mortality
in the unmatched cohort
Figure 10 - Prognostic value of fasting vs. non-fasting LDL-C level on cardiovascular mortality
in patients without diabetes in the unmatched cohort
Figure 11 - Prognostic value of fasting vs. non-fasting LDL-C level on cardiovascular mortality
in diabetic patients in the unmatched cohort
Figure 12 - Sensitivity analysis: Prognostic value of fasting vs. non-fasting LDL-C level
including those with triglycerides ≥400 mg/dL on cardiac mortality in the unmatched
cohort
Figure 13 - Sensitivity Analysis: Prognostic value of fasting (<4 hours) vs. non-fasting (≥4
hours) LDL-C level on cardiovascular mortality in the unmatched cohort
Figure 14 - Sensitivity Analysis: Prognostic value of fasting (<12 hours) vs. non-fasting (≥12
hours) LDL-C level on cardiovascular mortality in the unmatched cohort
Figure 15 - Prognostic value of fasting vs. non-fasting triglyceride level on cardiovascular
mortality in the unmatched cohort
Figure 16 - Prognostic value of fasting vs. non-fasting cholesterol level on cardiovascular
mortality in the unmatched cohort
3
Supplemental Figure 1 – Prognostic value of fasting vs. non-fasting LDL-C level
on all-cause mortality in the unmatched cohort
4
Supplemental Figure 2 – Sensitivity analysis: Prognostic value of fasting vs. non-
fasting LDL-C level including those with triglycerides ≥400 mg/dL on all-cause
mortality in the unmatched cohort
5
Supplemental Figure 3 – Prognostic value of fasting vs. non-fasting LDL-C level
on all-cause mortality in patients without diabetes in the unmatched cohort
6
Supplemental Figure 4 – Prognostic value of fasting vs. non-fasting LDL-C level
on all-cause mortality in diabetic patients in the unmatched cohort
7
Supplemental Figure 5 – Sensitivity Analysis: Prognostic value of fasting (<4
hours) vs. non-fasting (≥4 hours) LDL-C level on all-cause mortality in the
unmatched cohort
8
Supplemental Figure 6 - Sensitivity Analysis: Prognostic value of fasting (<12
hours) vs. non-fasting (≥12 hours) LDL-C level on all-cause mortality in the
unmatched cohort
9
Supplemental Figure 7 –Prognostic value of fasting vs. non-fasting triglyceride
level on all-cause mortality in the unmatched cohort
10
Supplemental Figure 8 –Prognostic value of fasting vs. non-fasting total
cholesterol level on all-cause mortality in the unmatched cohort
11
Supplemental Figure 9 – Prognostic value of fasting vs. non-fasting LDL-C level
on cardiovascular mortality in the unmatched cohort
12
Supplemental Figure 10 – Prognostic value of fasting vs. non-fasting LDL-C
level on cardiovascular mortality in patients without diabetes in the unmatched
cohort
13
Supplemental Figure 11 – Prognostic value of fasting vs. non-fasting LDL-C
level on cardiovascular mortality in diabetic patients in the unmatched cohort
14
Supplemental Figure 12 - Sensitivity analysis: Prognostic value of fasting vs.
non-fasting LDL-C level including those with triglycerides ≥400 mg/dL on
cardiovascular mortality in the unmatched cohort
15
Supplemental Figure 13 – Sensitivity Analysis: Prognostic value of fasting (<4
hours) vs. non-fasting (≥4 hours) LDL-C level on cardiovascular mortality in the
unmatched cohort
16
Supplemental Figure 14 – Sensitivity Analysis: Prognostic value of fasting (<12
hours) vs. non-fasting (≥12 hours) LDL-C level on cardiovascular mortality in the
unmatched cohort
17
Supplemental Figure 15 –Prognostic value of fasting vs. non-fasting triglyceride
level on cardiovascular mortality in the unmatched cohort
18
Supplemental Figure 16 –Prognostic value of fasting vs. non-fasting total
cholesterol level on cardiovascular mortality in the unmatched cohort
top related