The effects of almond consumption on fasting blood lipid levels

15 Pages·2016·1.78 MB·English

JNS
JOURNAL OF NUTRITIONAL SCIENCE
REVIEW ARTICLE
The effects of almond consumption on fasting blood lipid levels:
a systematic review and meta-analysis of randomised controlled trials
Kathy Musa-Veloso*, Lina Paulionis, Theresa Poon and Han Youl Lee
Intertek Scientific and Regulatory Consultancy, 2233 Argentia Road, Suite 201, Mississauga, Ontario, L5N 2X7, Canada
(Received7February2016–Finalrevisionreceived29April2016–Accepted9May2016)
JournalofNutritionalScience(2016),vol.5,e34,page1of15 doi:10.1017/jns.2016.19
Abstract
Asystematic reviewandmeta-analysisofrandomisedcontrolled trials wasundertakentodetermine theeffects ofalmond consumption onbloodlipid
levels, namely total cholesterol (TC), LDL-cholesterol (LDL-C), HDL-cholesterol (HDL-C), TAG and the ratios of TC:HDL-C and LDL-C:HDL-C.
Followingacomprehensivesearchofthescientificliterature,atotalofeighteenrelevantpublicationsandtwenty-sevenalmond-controldatasetswereiden-
tified.Acrossthestudies,themeandifferencesintheeffectforeachbloodlipidparameter(i.e.thecontrol-adjustedvalues)werepooledinameta-analysis
usingarandom-effectsmodel.ItwasdeterminedthatTC,LDL-CandTAGweresignificantlyreducedby−0·153mmol/l(P<0·001),−0·124mmol/l
(P=0·001)and−0·067mmol/l(P=0·042),respectively,andthatHDL-Cwasnotaffected(−0·017mmol/l;P=0·207).Theseresultsarealignedwith
datafromprospectiveobservationalstudiesandarecentlarge-scaleinterventionstudyinwhichitwasdemonstratedthattheconsumptionofnutsreduces
theriskofheartdisease.Theconsumptionofnutsaspartofahealthydietshouldbeencouragedtohelpinthemaintenanceofhealthybloodlipidlevels
andtoreducetheriskofheartdisease.
Keywords:Almonds:Bloodlipids:Cholesterol:TAG
Almonds are nutritionally dense(1). According to compositional meta-analysis of randomised controlled trials, Phung et al.(7)
datafromtheUnitedStatesDepartmentofAgriculture,100gof reported that almond consumption was associated with a
raw,unroastedalmondsprovides2423kJ(579kcal),50goffat, significant reduction in total cholesterol (TC) (−0·18mmol/
13g of insoluble dietary fibre and 21g of protein(2). There is l; 95% CI −0·34, −0·02mmol/l), as well as a strong trend
some natural variability in the compositionof almonds interms toward a reduction in LDL-cholesterol (LDL-C) (−0·15
ofthefatandfattyacidcontents;whenexpressedonaper100g mmol/l; 95% CI −0·29, 0·00mmol/l). No effects on
sserP basis, almonds contain about 45 to 54g of fat, with relative HDL-cholesterol (HDL-C), TAG, or on the ratio of LDL-
ytisrev a3mtoo5ungt,sreosfpePcUtivFeAly,(1M).IUnFadAdiatinodn,SaFlmAoonfds9cotont1a5in,s2m5atlolam36o,uanntds Cet:HalD.(7L)-Cwaswebraesedobosenrvfiedv.e Trahnedommeistead-ancaolynstirsollbedy sPtuhduinegs
inU e of plant sterols(3–5), and are sources or high sources of several (representing a total of 142 participants). More than 7 years
gdirbm mribinoeflraavlsina,nndiavciitnamanidnsvi(tia.em.CinaE,F),ea,cMcogr,dPin,gKt,oZthne,Creuq,uMiregm,tehniatmsfionr, hseaavreche;latphsuesd, aslilnrcaendPohmunisgedetcaoln.(7tr)oclloenddturciatlesdththaetirhalviteersaitnucree
a
C y nutritionclaims,assetoutinRegulationEC1924/2006(6). been published were identified, and, using the totality of
b
en Severalofthecompositional attributesofalmonds are ideal evidence, the effects of almonds on blood lipid levels were
iln
o d for the maintenance of healthy blood lipid levels.Indeed, in a re-evaluated.
e
h
silb
u
P
9
1
.6
10 Abbreviations:CAD,coronaryarterydisease;HDL-C,HDL-cholesterol;LDL-C,LDL-cholesterol;T2DM,type2diabetesmellitus;TC,totalcholesterol.
2
.snj/7 *Correspondingauthor:K.Musa-Veloso,fax[phone],[email protected]
1
0
1
.0
1
/gro ©TheAuthor(s)2016.ThisisanOpenAccessarticle,distributedunderthetermsoftheCreativeCommonsAttributionlicence(http://creative-
.io commons.org/licenses/by/4.0/),whichpermitsunrestrictedre-use,distribution,andreproductioninanymedium,providedtheoriginalworkis
d
//:sp properlycited. 1
tth
journals.cambridge.org/jns
Methods could not be isolated (e.g. almonds were co-consumed with
other nuts or another nutritional or pharmaceutical interven-
Literature search
tion); and (7) the study results for the same population
The systematic review was conducted in accordance with the group were published in another journal (i.e. the study was a
guidelines of the Preferred Reporting Items for Systematic kin publication to another study).
Reviews and Meta-analyses (PRISMA) statement.
In May 2013, the electronic search tool Dialog™ was used
to search eight literature databases (Allied & Complementary Data extraction and study quality
Medicine™, Adis Clinical Trials Insight, CAB Abstracts, Study data were extracted independently by two reviewers
Elsevier Biobase, EMBASE®, Foodline®: Science, Medline®, (T. P. and H. Y. L.), and the consistency of the two datasets
and National Technical Information Service). Dialog™ was was verified by a third reviewer (K. M. V.). Where there
subsequently replaced by ProQuest Dialog™, which did not were inconsistencies or discrepancies between the two data-
index one of the eight original literature databases that were sets, the original publication was consulted, and a consensus
searched (i.e. Elsevier Biobase). Updated literature searches was reached via discussions between the three reviewers
were conducted in May 2014 and in February 2015, using (T.P.,H.Y.L.andK.M.V.).Dataextractedfromthestudies
ProQuestDialog™toaccesssevenoftheeightoriginaldatabases. included study design, country of study conduct, sample size,
The search terms used reflected the exposure (‘almond’, study population (proportion of males, health status, mean
‘Prunus amygdalus’, ‘P. amygdalus’, ‘Prunus dulcis’, ‘P. dulcis’) and age, mean BMI, mean baseline blood lipid levels (i.e. TC,
the study population (‘human’, ‘subject’, ‘participant’, ‘volun- LDL-C,HDL-C,TAG,TC:HDL-C;LDL-C:HDL-C)),dietary
teer’, ‘patient’, ‘elder*’, ‘senior’, ‘geriatric’, ‘adult’, ‘men’, interventions (dose, form of almonds, provision of foods or
‘women’, ‘man’, ‘woman’, ‘teen*’, ‘adolescen*’, ‘people’, ‘per- meals, duration, frequency of intake, pattern of intake), back-
son’, ‘individual’), and were required to appear in the titles ground diets and macronutrient intakes, statistical results, and
or abstracts of the articles. The literature searches were the mean difference in the effect for each blood lipid param-
restricted to studies conducted in humans; that is, the key- eter (see the Statistical analysis section for details on how the
words used to limit the searches to human studies (‘animal’, mean difference in the effect was calculated for the crossover
‘rodent’, ‘rat’, ‘mouse’, ‘mice’, ‘dog’, ‘pig’, ‘rabbit’, ‘hamster’, and parallel studies).
‘monkey’, ‘in vitro’, ‘ex vivo’) were required not to appear in Health Canada’s standardised quality appraisal tool(8) was
the descriptor or subject field of the records. No restrictions used to determine the quality of the studies. A quantitative
with respect to the health outcomes of interest or language score (zero or one) was assigned to each of the fifteen items
were imposed on any of the literature searches. As well, no included in the tool, and studies with scores of ≥8/15 were
restriction on the year of publication was imposed on the considered to be ‘higher quality’ while studies with scores of
first literature search. ≤7/15 were considered to be ‘lower quality’. Study quality
was appraised by one reviewer (L. P.).
Inclusion/exclusion criteria
Statistical analysis
The following inclusion criteria were applied: (1) a human
intervention study that was randomised and controlled (such Several of the identified studies had multiple comparisons
that the control group/phase could have consisted of either (e.g.onestudymayhavehadthreearms,includingonecontrol
no food or any food(s) without tree nuts or fractions of tree and two different almond doses). Each almond-control com-
nuts); (2) a full-length article that was published in a peer- parison, hereinafter referred to as a stratum, was considered
sserP roervsieewcoedndjoaruyr)nwala;s(3t)othaessoesbsjetchtieveefofefctthseosftuadlmyo(enitdhecropnrsiummapry- aevseenplyaraatmeotrnigasl;thtohweecvoemr,ptahreisoconnstrsoolassamtoplaevsoiizdeiwnflasatdinivgidtehde
y
tisrevin tsiuomneodnwbalsooredpoliprtiedd;le(v5e)lst;h(e4)sutbhjeecatms wouenret aodfualtlms (oangdesdc≥o1n8- wenecigehint othfeeeafcfhecsttrfoatruemac.hFbolroopdaralilpleidl sptaurdaimese,tethrewmasecaanlcduilfafteerd-
U
e years of age) withoutseriousdiseasesuchasheartdisease;(6) as the change from baseline in the control group subtracted
g
dirb the study duration was ≥4 weeks; (7) fasting blood lipids (i.e. from the change from baseline in the almond group. For
m
a TC,LDL-C,HDL-Cand/orTAG)wereassessed;and(8)fast- crossover studies, the mean difference in the effect for each
C
yb ingbloodlipidsweremeasuredusingvalidatedmethods. blood lipid parameter was calculated as the blood lipid value
e
niln The following exclusion criteria were applied: (1) the publi- at the end of the control phase subtracted from the blood
o
d cation was of a secondary research study (e.g. systematic re- lipid value at the end of the almond phase.
e
h
silb view or meta-analysis); (2) the objective of the study (either Inordertodeterminetheeffectsofalmondsoneachofthe
uP primaryorsecondary)wasnottoassesstheeffectsofalmond blood lipid parameters, the results of the studies were pooled
9
1.6 consumption on blood lipid levels; (3) the subjects had a ser- in a meta-analysis, with the mean difference in the effect and
1
02 iousdisease(e.g.heartdisease,cancer)andwerenotrepresen- theinverseofthevarianceusedastheweightingfactor.Inthe
.sn
j/7 tativeof thegeneralpopulation;(4) the subjects werechildren majority of the studies, variances for the mean differences
1
01 orpregnantorlactatingwomen;(5)control-adjustedeffectson werenotreported;thus,varianceswerecalculatedusinginfor-
.0
1/g blood lipids could not be calculated from the data provided; mationprovidedinthepublication(e.g.usingCIorindividual
ro.io (6) the independent effects of almonds on blood lipid levels variances for the almond and control groups). If, in parallel
d
//:sp 2
tth
journals.cambridge.org/jns
studies,variancesforthechangesfrombaselinewerereported of the strata, with the exception of three strata wherein only
separately for the almond and control groups, then a pooled females were studied (Abazarfard et al.(9); Kurlandsky &
varianceforthemeandifferencewascalculated.If,forparallel Stote strata 1 and 2(14)) and two strata wherein only males
studies, variances only for the baseline and end of treatment were studied (Jia et al. strata 1 and 2(15)). The subjects were
values were reported, then these were used to calculate the described by the authors as generally healthy in seven strata
variance for the change from baseline, using a correlation co- (Abazarfard et al.(9); Berryman et al.(11); Foster et al.(16);
efficient of 0·8. Similarly, for crossover studies, if variances Kurlandsky & Stote strata 1 and 2(14); Spiller et al. strata 1
only for the end of treatment values were reported, then the and2(17)),generallyhealthybuthabitualsmokersintwostrata
varianceforthemeandifferencewascalculatedusingacorrel- (Jiaetal.strata1and2(15)),generallyhealthyorhyperlipidaemic
ation coefficient of 0·8. A correlation coefficient of 0·8 was intwostrata(Sabatéetal.strata1and2(18)),hyperlipidaemicin
used because this value approximated that calculated from fourstrata(Damascenoetal.(19);Jenkinsetal.strata1and2(20);
the studies in which variances were provided for the baseline, Tamizifaretal.(21)),orpre-diabeticoratriskoftype2diabetes
end of treatment, and change from baseline measures(9–11). mellitus (T2DM) in fivestrata(Tan & Mattesstrata1 to 4(22);
A random-effects model was used, according to the methods Wien et al.(23)). In the remaining seven strata, medicated
described by DerSimonian & Laird(12), given that random- subjects were studied, including subjects with T2DM on
effects models take into consideration the variability in re- oral hypoglycaemic therapy (Sweazea et al.(10); Cohen &
sponse both within and between studies. Johnston(24); Li et al.(25); Lovejoy et al. strata 1 and 2(26)),
The pooled estimates and accompanying 95% CI were subjects on stable statin therapy (Ruisinger et al.(27)), and sub-
determinedusingComprehensiveMeta-analysisSoftware(ver- jects on antihypertensive and/or antidiabetic medications
sion 2.2.064). Publication bias was assessed according to the (Wien et al.(28)).
trim-and-fill method developed by Duval & Tweedie(13).
With this method, asymmetry in the funnel plot is searched
for.Iftheasymmetryisdetermined tobedue tothepresence
Almond interventions
of small studies (with large variances) in which large effect
sizeswerereported,withanunbalancednumberofsmallstud- Across all strata, the average daily intake of almonds ranged
ies showing a small effect, then those ‘missing’ studies are from20to113g/d,andthedurationofthealmondconsump-
imputed, and the pooled effect size is recalculated. tionperiodrangedfrom4weeksto18months.Almondswere
Subgroup analyses were conducted to evaluate the influence requiredtobeconsumedeverydayinallstudiesexcepttwo,in
ofdose(i.e.<45v.≥45g/d),studydesign(i.e.parallelorcross- which28g(1oz)ofalmondswererequiredtobeconsumed5d
over), the control food/diet (i.e. whether it was provided or if per week(24) or 43g (1·5oz) of almonds were required to be
subjects were simply instructed to avoid nuts), the duration of consumed five to seven times weekly(10).
thestudy(i.e.≥12weeksv.4to<12weeks(hereinafterreferred Whole,raw(unblanched,unsalted)almondswereconsumed
to as <12 weeks)), and of baseline blood lipid level. Baseline inninestrata(Abazarfardetal.(9);Sweazeaetal.(10);Damasceno
blood lipid levels were categorised dichotomously as ‘optimal’ et al.(19); Jenkins et al. strata 1 and 2(20); Ruisinger et al.(27);
or ‘not optimal’, based on the targets established in the Spiller et al. strata 1 and 2(17); Wien et al.(28)). In five strata
National Cholesterol Education Program Adult Treatment (Cohen & Johnston(24); Tan & Mattes strata 1 to 4(22)), the
PanelIIIguidelines(i.e.optimalbloodlipidlevelsweredefined almondsthatwereconsumed by the subjects were not specif-
as: LDL-C<2·59; TC<5·17; HDL-C≥1·03; TAG<1·69 ically described by the study authors as raw, unblanched
mmol/l). For crossover studies, the categorisation was based almonds; however, based on the reported energy value of
onthereportedbaselinelipidlevel;forparallelstudies,thecat- the almonds, it was determined that the almonds were most
sserP ethgaotriwsaetrieonrewpoasrtbedasfeodroenacthhegraovuepra,gweeoigfhthteedbbaysetlhineesalimpidplelevsiezles pnraombealbylywhroawle,(ruanwb),larnocahsteedd,aalmndoflndavs.ouAredvaarlmietoyndosfwaelmreocnodns-,
y
tisre ofeachgroup.Subgroupanalyseswereconductedwhenthere sumed in one stratum (Foster et al.(16)); dry, roasted almonds
vin were at least three strata available for pooling. were consumed in one stratum (Wien et al.(23)); and almond
U
e powder was consumed in one stratum (Tamizifar et al.(21)).
g
dirb The types of almonds used were not specified in two strata
m
aC Results (Kurlandsky & Stote strata 1 and 2(14)). In eight strata, all
yb meals and snacks were provided and the almonds were said
eniln Literature search results and overview of included studies to have been consumed as a snack (Berryman et al.(11) or
o d The three literature searches resulted in the identification of incorporated into the meals and snacks (Jia et al. strata 1 and
e
hsilb 1697titles,ofwhicheighteenpublicationsmetalloftheinclu- 2(15); Li et al.(25); Lovejoy et al. strata 1 and 2(26); Sabaté et al.
uP sion criteria and none of the exclusion criteria (Fig. 1). strata 1 and 2(18)). The form of almonds that was used was
9
1.6 The eighteen publications provided a total of twenty-seven describedonlybyBerrymanetal.(11),whoreportedadminister-
1
02 strata (Table 1). Of the twenty-seven strata, seventeen were ing unsalted, whole, natural almonds with skins, and by Jia
.snj/7 from parallel trials, and the remaining ten strata were from et al. strata 1 and 2(15), who reported using almond powder.
101 crossover trials. The number of study completers ranged In the remaining five strata in which all meals and snacks
.01/g from thirteen to 137 amongst the eighteen publications. were provided (Li et al.(25); Lovejoy et al. strata 1 and 2(26);
ro.io Both male and female subjects were studied in the majority Sabaté et al. strata 1 and 2(18)), it is assumed that whole
d
//:sp 3
tth
journals.cambridge.org/jns
Fig.1. Flowchartoftheliteraturesearchprocess.
almonds, almond pieces and ground almonds were used to Study quality
prepare the meals. Based on Health Canada’s quality appraisal tool, all of the
studies were considered to be ‘higher quality’(8). Across all
sserP eighteen publications, the most commonly identified limita-
ytisrevin CAoltnhtoroulgfhooaldlss/tduideitesswererandomisedandcontrolled,thecon- t(nion1s6)i,ntchluedleadckthoeflraecpkoortfinrgepoonrtitnhgeomneathlloodcaotifonracnodnocmeaislmatieonnt
U e trolfoodwasnotdefinedinsomestudiesbutdefinedinother and thus the ‘appropriateness’ of the randomisation method,
g
dirb studies. In thirteen of the twenty-seven strata, subjects in the which also is a quality factor, could not be determined
ma control group or during the control phase were instructed (n 13), as well as the lack of reporting of an intent-to-treat
C
yb not to consume nuts, but were not provided with a control analysis (n 13).
e
niln food or with a control diet (Abazarfard et al.(9); Sweazea et al.(10);
o d Foster et al.(16); Kurlandsky & Stote strata 1 and 2(14);
ehsilb Ruisinger et al.(27); Tamizifar et al.(21); Tan & Mattes strata 1 Effects of almonds on fasting blood lipids
uP to4(22);Wienetal.(28);Wienetal.(23)).Infourteenstrata,either The fasting blood lipids that were assessed in each of the
9
1.6 a‘controlfood’(e.g.cheesesticksoramuffinoroliveoil)was twenty-seven strata, as well as other information pertinent to
1
02 provided(Cohen&Johnston(24);Damascenoetal.(19);Jenkins the subgroup analyses (i.e. study design, almond intake,
.snj/7 et al. strata 1 and 2(20); Spiller et al. strata 1 and 2(17)), or the whether a control food/diet was provided, study duration,
101 entire control diet was provided (Berryman et al.(11); Jia et al. and baseline blood lipid levels), are summarised in Table 2.
.01/g strata 1 and 2(15); Li et al.(25); Lovejoy et al. strata 1 and 2(26); In the study by Lovejoy et al. strata 1 and 2(26), baseline
ro.io Sabaté et al. strata 1 and 2(18)). TAG levels were not reported; thus, a determination as to
d
//:sp 4
tth
https://doi.org/10.1017/jns.2016.19 Published online by Cambridge University Press
Table1. Keystudycharacteristicsofincludedstudies(n18publicationsand27strata)
jo
u
Meanbaseline r
n
a
Studypopulation ls
References Studydesign Studyduration (finalsamplesize) Age(years) BMI(kg/m2) Control Almondintervention Provisionofcontrolfoods/diet .c
a
m
Abazarfard R,C,P 12weeks 100F;generallyhealthy; 42·7±7·1* 29·6±1·5* Nonuts(n50) 50g/dalmonds(raw) Hypoenergeticdiet b
r
etal.(9) non-medicated astwosnacks(about prescribed;foodswere id
25g/snack)(n50) self-selectedfroma ge
providedlist .o
Berryman R,C,X 6weeks 48(22M,26F);generally 49·9 ± 9·4 26·2 ±2·8 106g/dbanana 42·5g/d(1·5oz) Mealspreparedbya rg
etal.(11) (2weeksWO) healthy;non-medicated muffinwith2·7g/d wholealmonds metabolickitchen /jn
butter (unsalted,unroasted) s
Cohen& R,C,P 12weeks 13(7M,6F);T2DM; 66 ±8·4* 34·8 ±8·0* Twocheesesticks, 20g/dalmonds(28g Foodswereself-selected
Johnston(24) medicated† 5d/week(n7) or1oz,5d/week)(n6)
Damasceno R,C,X,SB§ 4weeks 18(9M,9F); 56 ±13* 25·7 ±2·3* 35to50g/doliveoil 50to75g/dalmonds Mediterranean-typediet
etal.(19)‡ (noWO) hypercholesterolaemic; (raw,shelled,Spanish prescribed;foodswere
non-medicated Marconavariety)║ self-selected¶
Fosteretal.(16) R,C,P 18months 92(MandF);generally 46·8 ± 12·5* 34·0 ±3·6* Nonutsorpeanut 56g/dalmonds(n47) Low-energydietprescribed;
healthy;non-medicated butter(n45) ** foodswereself-selected
Jenkinsetal. R,C,X 4weeks(≥2 27(15M,12F); 64 ±9 25·7 ±3 147±6g/dmuffin 37 ±2g/dwhole Self-selectedlow-fat
stratum1(20) weeksWO) hyperlipidaemic; almonds(raw, therapeuticdiet(selected
non-medicated†† unblanched)with followingdietaryinstruction)
75 ±3g/dmuffin
Jenkinsetal. 73 ±3g/dwhole
stratum2(20) almonds(raw,
unblanched)
Jiaetal. R,C,P 4weeks 30M;generallyhealthy, 22·3 ± 1·8* NR Noalmonds(n10) 84g/d(3oz)almond Mealsprovidedbyarmyunit
stratum1(15) habitualsmokers;partofan powder(n10) canteen
armyunit;medicationstatus
NR
Jiaetal. 22·1 ± 1·8* 168g/d(6oz)almond
stratum2(15) powder(n10)
Kurlandsky& R,C,P 6weeks 47F;healthy 46·6 ± 9·4* 25·7 ±3·8* Nonutsor 60g/dalmonds(n12) Darkchocolateand
Stote normocholesterolaemic; chocolate(n12) almondsprovided;foods
stratum1(14) non-medicated‡‡ wereotherwise
self-selected
Kurlandsky& 41·1 ± 10·2* 25·5 ±3·8* 41g/ddark 41g/ddarkchocolate
Stote chocolate(n12) and60g/dalmonds
stratum2(14) (n11)
Lietal.(25) R,C,X 4weeks 20(9M,11F);T2DM; 58 ±2 26·0 ±0·7 Noalmonds About56g/dwhole Mealspreparedby
(2weeksWO) medicated§§ almonds(roasted, metabolickitchen
unsalted,
unblanched);20%of
TDEIfromalmonds║║
Lovejoyetal. R,DB,X¶¶ 4weeks 30(13M,17F);T2DM; 53·8 ± 10·4 33·0 ±1·0 Oliveoilorcanola 57to113g/dalmonds Mealsprovided†††
stratum1(26) (2weeksWO) medicated*** oil;fatcomprised (inmealsandsnacks);
25%ofTDEI,with fatcomprised25%of
10%fromoliveor TDEI,with10%from
canolaoil almonds
Continued
5
https://doi.org/10.1017/jns.2016.19 Published online by Cambridge University Press
Table1. Continued
jo
u
Meanbaseline rn
a
Studypopulation ls
References Studydesign Studyduration (finalsamplesize) Age(years) BMI(kg/m2) Control Almondintervention Provisionofcontrolfoods/diet .ca
m
Lovejoyetal. Oliveoilorcanola 57to113g/dalmonds b
r
stratum2(26) oil;fatcomprised (inmealsandsnacks); id
g
37%ofTDEI,with fatcomprised37%of e
10%fromoliveor TDEI,with10%from .o
r
canolaoil almonds g
/
Ruisinger R,C,P 4weeks 48(24M,24F);medicated(on 59·6 ± 11·1* 29·2 ±4·3* NCEPATPIIIdiet NCEPATPIIIdiet Dietcounselling jn
etal.(27) stablestatintherapy‡‡‡) counselling(n26) counsellingand100g/ provided§§§ s
dwholealmonds(raw,
unsalted)(n22)
Sabatéetal. R,C,X 4weeks 25(14M,11F);generally NR(20to60) NR¶¶¶ Noalmonds About34galmonds/ Mealspreparedbya
stratum1(18) (noWO║║║) healthyormildly 2000kcal(inmeals metabolickitchen****
hypercholesterolaemic; andsnacks);10%of
non-medicated TDEIfromalmonds
Sabatéetal. About68galmonds/
stratum2(18) 2000kcal(inmeals
andsnacks);20%of
TDEIfromalmonds
Spilleretal. R,C,P 4weeks 45(12M,33F);generally 53 ±10†††† NR 48g/doliveoil,113 100g/dwholeor Dietwasbothprescribed
stratum1(17) healthy;non-medicated g/dcottagecheese, groundalmonds(raw, andself-selected‡‡‡‡
and21g/drye unblanched)(n18)
crackers(n15)
Spilleretal. 85g/dcheddar
stratum2(17) cheese,28g/d
butter,and21g/d
ryecrackers(n12)
Sweazea R,C,P 12weeks 21(9M,12F);T2DM; 56·2 ± 7·5* 35·3 ±8·3* Noalmonds(n11) 43g(1·5oz)whole Foodswereself-selected
etal.(10) medicated§§§§ almonds,fivetoseven
times/week(n10)
Tamizifar R,SB║║║║, 4weeks 30(17M,13F); 56 ±6·1 24·1 ±4·5 Nonuts,nutbutter 25g/dalmondpowder Foodswereself-selected
etal.(21) C,X (5to7dWO) hypercholesterolaemic; ormargarines,or fromaprovidedlist
non-medicated nutoils
Tan&Mattes R,C,P 4weeks 137(48M,89F);atriskof 30·8 ± 10·6* 27·6 ±4·6* Nonutsorseeds 43g/dalmondswith Foodswereself-selected
stratum1(22) T2DM;non-medicated¶¶¶¶ (n27) breakfast(n28)
Tan&Mattes 28·2 ± 10·2* 27·9 ±4·7* 43g/dalmondsas
stratum2(22) morningsnack(2h
afterbreakfastand2h
beforelunch)(n28)
Tan&Mattes 29·0 ± 11·7* 28·0 ±4·2* 43g/dalmondswith
stratum3(22) lunch(n26)
Tan&Mattes 28·9 ± 10·8* 27·6 ±4·8* 43g/dalmondsas
stratum4(22) afternoonsnack(2h
afterlunchand2h
beforedinner)(n28)
6
https://doi.org/10.1017/jns.2016.19 Published online by Cambridge University Press
Wienetal.(28) R,C,P 24weeks 52(MandF);withamedical 55±2·0* 38·0 ±1·0* Low-energyliquid Low-energyliquid Low-energyliquidformula jo
diagnosisthatcouldbenefit formula+complex formula+84g/d provided;complexCHO u
fromweightreduction***** CHO(n28) whole,unblanched foodswereself-selected rn
a
almonds fromaprovidedlist ls
(n24) .c
Wienetal.(23) R,C,P 16weeks 54(MandF); 53·5 ± 10·1* 29·5 ±5* Notreenutsor About60g/dalmonds Foodswereself-selected a
m
prediabetic†††††; peanuts(n29) (rawordryroasted); b
non-medicated‡‡‡‡‡ 20%ofTDEIfrom rid
almonds(n25) g
e
.o
R,randomised;C,controlled;P,parallel;F,females;X,crossover;WO,washout;M,males;T2DM,type2diabetesmellitus;SB,single-blinded;NR,notreported;TDEI,totaldailyenergyintake;DB,double-blinded;NCEPATPIII,National r
CholesterolEducationProgramAdultTreatmentPanel’sThirdReportonTherapeuticLifestyleChanges;CHO,carbohydrates;ACE,angiotensin-convertingenzyme;HRT,hormonereplacementtherapy;ADA,AmericanDiabetesAssociation. g/
*ForPstudies,theweightedmeanage/BMIandpooledSDwerecalculatedbyusingthemeanage/BMIandSDvaluesthatwerereportedseparatelyfortheindividualgroups. jn
†Subjectshadnotbeenprescribedinsulin;rather,subjectswereonstableoralhypoglycaemictherapy. s
‡Inanotherarmofthistrial,subjectswereadministered40to65g/dofwalnuts;however,accordingtothestudyexclusioncriteria,thecontrolisnotacceptableifitconsistsofanothertreenutoranytreenutfraction.Thus,theresultsofthewalnuts
interventionarmwereexcludedfromtheanalysis.
§Damascenoetal.(19)acknowledgedthat,giventhenatureofthefoodsprovided,whichcouldnotbemasked,thestudywasunblinded;however,itwasnotedthatinvestigatorsinvolvedinthepreparationofdatabasesandlaboratorydeterminations
weremaskedwithrespecttotreatmentsequence.
║ThealmondspartiallyreplacedotherMUFA-richfoods,suchasolivesandavocados.
¶TheMediterranean-type,cholesterol-loweringdietwascomposedofnaturalfoodstuffs.Vegetableproductsandfishwereemphasised,whileredandprocessedmeats,whole-fatdairyproductsandeggswerelimited.Recipesusingnutswere
providedtothesubjectswhoconsumedthenutswithmealsindessertsorsaladsorassnacks.
**Duringthefirst5weeks,subjectsreceivedwhole,rawalmondsonly.Fromweek6,roastedalmondswereintroducedand,overtime,avarietyofisoenergetic,flavouredalmondswereused.
††Themajorityofthesubjectswerenotmedicated;however,3Mand5Fweretakingthefollowingmedications:statins(n2),β-blockingagents(n3),ACEinhibitors(n3),angiotensinIIAT1receptorblockers(n1),thiazidediuretics(n2),
levothyroxine(n2)andHRT(n2).Medicationdosageswereheldconstantthroughoutthestudy.
‡‡Subjectsdidnotuselipid-loweringmedicationsordietarysupplements.SubjectswereallowedtobeonstableregimensoforalcontraceptivesandHRT.
§§Subjectsdidnotreceiveinsulintherapy;rather,allsubjectswereonstableoralhypoglycaemictherapy.
║║Almondswereincorporatedintothecontroldiettoreplace20%ofTDEIinthecontroldiet;dependingonthemenus,almondswereeitherincorporatedintoentréesanddessertsorconsumedasasnack.
¶¶Inthisstudy,subjectswereprovidedwithallfoodsneededforthedurationofthestudy;sincethestudywascrossoverindesignbutdescribedasDB,itisassumedthatalmondpowderwasusedintheformulationoftheentréesandsnacks.
However,itremainsunclearhowtruedouble-blindingwasachieved,giventhatthesubjectscouldhavetastedthealmondsintheirfoods(itispossiblethatthe‘control’foodswereflavouredwithanalmondextract;however,thiswasnotdescribedin
thepublication).
***Subjectsonlipid-loweringorinsulintherapywereexcluded.ElevenFreceivedHRT,andsixteenofthesubjectsweretakingoralhypoglycaemicagents.
†††Regardingbackgrounddiets,onweekdays,thesubjectswererequiredtoconsumebreakfastanddinnerundersupervisionatthePenningtonBiomedicalResearchCenter’sdiningfacility;weekdaylunchesandsnacksandallweekendmeals
werepackagedfortake-out.
‡‡‡Subjectsweretakingchronicstatintherapy,definedasaconsistentstatindoseforatleast8weeksbeforestudyentrywithcontinuationofthesamedoseduringthe4-weekstudyperiod.Subjectswhotooklipid-loweringagentsotherthan
statinswereexcluded.Post-menopausalFwhowerenottakingHRTorwereonaconsistentHRTdosewereincluded.Fofchild-bearingpotentialusinganeffectiveformofcontraceptionwereallowedtoparticipateinthestudy.
§§§SubjectsreceivedNCEPATPIIIdietcounsellingviatelephoneandinstructionsonhowtocompensatefortheaddedenergyfromthealmonds.
║║║TheauthorsarguedthataWOwasnotincludedbecauseserumlipidsandlipoproteinsareknowntostabilisewithin3weeks.
¶¶¶AlthoughthemeanBMIwasNR,itshouldbenotedthataBMI>30kg/m2wasanexclusioncriterion.
****OnSundaytoFridayofeachweek,subjectsatebreakfastanddinnerattheLomaLindaUniversitymetabolickitchen.LunchmealsandallSaturdaymealswerepackagedforconsumptionawayfromthemetabolickitchen.
††††Themeanagewasprovidedfortheforty-fivestudycompleters.Becausethemeanagewasnotprovidedforeachgroup,theweightedmeanageforeachcomparisoncouldnotbecalculated.
‡‡‡‡Allsubjectswereprovidedwithwhole-grainbread,brownrice,pasta,non-fatyogurt,ricecakes,drybeans,lentils,andcouscousandwereinstructedtoeatthesefoodsasetnumberoftimesduringeachweek.Subjectsroundedouttheirdaily
foodintakewithfruits,vegetables,otherwholegrains,legumes,low-fatornon-fatmilk(wholemilkdairyproductswerenotallowed),eggwhitesandleanfish.Leanbeefwasallowedtwiceweekly,andpoultryandfattyfishwerepermitteduptofour
timesweekly.Uptofourwholeeggswerepermitted/week,butonlyifthesubjecthadbeenconsumingeggspriortothestudy.Foodsnotallowedincluded:commercialorhomemadeproductscontainingfatsotherthanthestudyfat,andproducts
madewithrefinedflour(e.g.snackfoods,chips,crackers,cakes,pastries,pies,candyoricecream).Usualcoffee,tea,alcoholandsoftdrinkconsumptionwaspermitted.
§§§§Subjectsoninsulintherapywereexcluded.Subjectstakingprescriptionmedications,includingoralhypoglycaemicagents,statinsorhypertensivemedicationswereinstructedtomaintainconsistentusethroughoutthestudy.
║║║║ThestudybyTamizifaretal.(21)wasdescribedasSB.Giventhat,duringthealmondphaseofthecrossoverstudy,thesubjectsweregivenalmondpowderand,duringthecontrolphase,thesubjectswerenotadministeredanything,itis
assumedthattheinvestigatorswereblinded,andnotthesubjects.
¶¶¶¶SubjectswereconsideredatriskofT2DMiftheywereoverweightorobese(BMI>27kg/m2)orwerenormalweight(BMI18·5–24·9kg/m2)buthadastrongfamilyhistoryforT2DM.
*****Theproportionofsubjectswithadiagnosisofhypertensionatbaselinewas62%inthecontrolgroupand50%inthealmondgroup;theproportionofsubjectswithT2DMwasNR.However,itwasnotedthatduringrandomisation,subjects
werestratifiedaccordingtothepresenceorabsenceofT2DM.
†††††Prediabeteswasdiagnosedaccordingtothe2005ADAdiagnosticguidelines:fastingbloodglucosebetween100and125mg/dl(5·56and6·94mmol/l)orcasualbloodglucose≥140–199mg/dl(≥7·78–11·06mmol/l).
‡‡‡‡‡Subjectstakingcorticosteroidsorimmunosuppressantmedicationswereexcluded.Twosubjectsineachgroupweretakinglipid-loweringmedications.
7
https://doi.org/10.1017/jns.2016.19 Published online by Cambridge University Press
jo
u
r
n
a
ls
.c
a
m
Table2. Summaryofstudydesignandduration,almonddoseandbaselinebloodlipids b
r
id
Study Controlfood/ Duration g
e
design Dose(g/d) diet (weeks) TC LDL-C HDL-C TAG .o
r
g
References P X ≥45 <45 Pr NPr <12 ≥12 BL* EOT BL* EOT BL* EOT BL* EOT TC:HDL-C LDL-C:HDL-C /
jn
s
Abazarfardetal.(9) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ N.O. ✓ ✓ –
Berrymanetal.(11) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ ✓ ✓
Cohen&Johnston(24) ✓ ✓ ✓ ✓ O ✓ O ✓ – – N.O. ✓ – –
Damascenoetal.(19) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ – ✓
Fosteretal.(16) ✓ ✓ ✓ ✓ O ✓ N.O. ✓ O ✓ O ✓ ✓ –
Jenkinsetal.stratum1(20) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ N.O. ✓ ✓ ✓
Jenkinsetal.stratum2(20) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ ✓ ✓
Jiaetal.stratum1(15) ✓ ✓ ✓ ✓ O ✓ – – – – O ✓ – –
Jiaetal.stratum2(15) ✓ ✓ ✓ ✓ O ✓ – – – – O ✓ – –
Kurlandsky&Stotestratum1(14) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ – –
Kurlandsky&Stotestratum2(14) ✓ ✓ ✓ ✓ O ✓ N.O. ✓ O ✓ O ✓ – –
Lietal.(25) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ – ✓
Lovejoyetal.stratum1(26) ✓ ✓ ✓ ✓ O ✓ N.O. ✓ O ✓ – ✓ ✓ ✓
Lovejoyetal.stratum2(26) ✓ ✓ ✓ ✓ O ✓ N.O. ✓ O ✓ – ✓ ✓ ✓
Ruisingeretal.(27) ✓ ✓ ✓ ✓ O ✓ N.O. ✓ O ✓ N.O. ✓ – –
Sabatéetal.stratum1(18) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ – ✓
Sabatéetal.stratum2(18) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ – ✓
Spilleretal.stratum1(17) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ –† –† –† –† – –
Spilleretal.stratum2(17) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ –† –† –† –† – –
Sweazeaetal.(10) ✓ ✓ ✓ ✓ O ✓ N.O. ✓ O ✓ N.O. ✓ – –
Tamizifaretal.(21) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ N.O. ✓ N.O. ✓ ✓ –
Tan&Mattesstratum1(22) ✓ ✓ ✓ ✓ O ✓ O ✓ O ✓ O ✓ – –
Tan&Mattesstratum2(22) ✓ ✓ ✓ ✓ O ✓ O ✓ O ✓ O ✓ – –
Tan&Mattesstratum3(22) ✓ ✓ ✓ ✓ O ✓ O ✓ O ✓ O ✓ – –
Tan&Mattesstratum4(22) ✓ ✓ ✓ ✓ O ✓ O ✓ O ✓ O ✓ – –
Wienetal.(28) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ N.O. ✓ N.O. ✓ – ✓
Wienetal.(23) ✓ ✓ ✓ ✓ N.O. ✓ N.O. ✓ O ✓ O ✓ ✓ –
Total 17 10 17 10 14 13 21 6 14N.O. 27 20N.O. 25 2N.O. 22 7N.O. 25 9 10
13O 5O 20O 16O
TC,totalcholesterol;LDL-C;LDL-cholesterol;HDL-C;HDL-cholesterol;P,parallel;X,crossover;Pr,provided;NPr,notprovided;BL,baseline;EOT,endoftreatment;N.O.,notoptimal;O,optimal;–,notreported.
*MeanbaselineTC,LDL-C,HDL-CandTAGwerecategorisedasOorN.O.,basedonthetargetsestablishedintheNationalCholesterolEducationProgramAdultTreatmentPanelIIIguidelines(i.e.optimalbloodlipidlevelsweredefinedas:
TC<5·17mmol/l;LDL-C<2·59mmol/l;HDL-C≥1·03mmol/l;TAG<·69mmol/l).
†InthestudybySpilleretal.strata1and2(17),HDL-CandTAGlevelswereassessedatBLandatEOT;however,valueswerepresentedonlyinfigureform,withnomeasuresofvariability.TheresultsrelatedtoHDL-CandTAGcouldnotbe
includedinthemeta-analyses.
8
journals.cambridge.org/jns
whether TAG levels at baseline were or were not optimal in which all twenty-two strata were pooled or in any of the
could not be made. In the study by Spiller et al. strata 1 and subgroup analyses (Table 3 and Fig. 4).
2(17),TC,LDL-C,HDL-CandTAGwereassessedatbaseline The effects of almonds on fasting TAG were assessed in
and atthe end of treatment; however, only the results for TC twenty-five strata. The almond intake was 45g or greater in
and LDL-C at the end of treatment were reported. Thus, the 60% of the strata, the design was crossover in 40% of the
HDL-C and TAG results could not be included in the strata, the baseline fasting TAG level was not optimal
meta-analyses, and in order to include the results related to in 30% of the strata, a control food/diet was provided in
TC and LDL-C in the meta-analyses, the mean difference in 48% of the strata, and the study duration was <12 weeks in
theeffectforTCandLDL-Chadtobecalculatedbysubtract- 76% of the strata (Table 3). As can be seen in Table 3 and
ing the end-of-treatment value in the control group from the Fig. 5, the reduction in TAG was statistically significant
end-of-treatment value in the almond group (as opposed to when data from all twenty-five strata were pooled (−0·067
subtracting the changes from baseline in the control group mmol/l; 95% CI −0·132, −0·002mmol/l; P=0·042). As
from the changes from baseline in the almond group, which there was no publication bias identified, no adjustment to
was done for all other parallel studies). these values was made. In all of the subgroup analyses, the
TheeffectsofalmondsonfastingTClevelswereassessedin pooledeffect sizesfor TAGwerenegativebut notstatistically
all twenty-seven strata. The daily almond intake was 45g or significant, except for when the fifteen strata in which the
greater in 63% of the strata, the design was crossover in study design was parallel were pooled, and the resultant
37% of the strata, the baseline fasting TC level was not opti- pooled effect was a statistically significant reduction in TAG
malin52%ofthestrata,acontrolfood/dietwasprovidedin (−0·111mmol/l; 95% CI −0·204, −0·017; P=0·020).
52% of the strata, and the study duration was <12 weeks in Through additional sensitivity analyses, it was determined
78% of the strata (Table 3). As can be seen in Table 3 and that this effect was dependent on the inclusion of the parallel
Fig. 2, the reduction in TC was statistically significant when study by Abazarfard et al.(9), which included 100 females and
data from all twenty-seven strata were pooled (−0·153 was a relatively larger study.
mmol/l; 95% CI −0·235, −0·070mmol/l; P<0·001). As WithregardstotheratioofTC:HDL-C,whendatafromall
there was no publication bias identified, no adjustment to ninestratawerepooled,thereductionintheratiowasstatistic-
these values was made. In all of the subgroup analyses, the ally significant (see Table 3 and Fig. 6). However, publication
pooledeffectsizesforTCwerenegative.Statisticalsignificance biaswasdetected. Using trim and fill, twostudieswere found
was observed when pooling those strata in which the almond to be missing to the right of the pooled effect size, and with
dosewas≥45g/d, the baseline TClevelwasnot optimal, the these studies imputed, the pooled effect, though negative
studydesignwaseitherparallelorcrossover,thecontrolfood/ (i.e.favourable),wassmallerandnolongerstatisticallysignificant
diet either was or was not provided, and the duration of the (Table 3). Results for the subgroup analyses were in the same
almond intervention period was <12 weeks (Table 3). directionofeffect(i.e.thepooledeffectwasnegative),withvari-
The effects of almonds on fasting LDL-C were assessed in able statistical significance. With regards to the ratio of LDL-C:
twenty-fivestrata.Thedailyalmondintakewas45gorgreater HDL-C,whendatafromalltenstratawerepooled,thereduction
in60%ofthestrata,thedesignwascrossoverin40%ofthe intheratiowasnotsignificant(−0·089;95%CI−0·209,0·031;
strata,thebaselinefastingLDL-Clevelwasnotoptimalin80 P=0·145) (see Table 3 and Fig. 7). As for the ratio of TC:
% of the strata, a control food/diet was provided in 48% of HDL-C, publication bias was detected, and using trim and fill,
the strata, and the study duration was <12 weeks in 76% twostudieswerefoundtobemissingtotherightofthepooled
of the strata (Table 3). As can be seen in Table 3 and effect size for LDL-C:HDL-C. With these studies imputed, the
Fig. 3, the reduction in LDL-C was statistically significant pooled effect, thoughnegative (i.e. favourable),was smallerand
sserP wmhmeonl/dl;ata95fr%omCIall−tw0·e1n9t6y,-fi−ve0·s0t5ra1tamwmeorle/lp;oPo=led0·(0−010)·.12A4s rseumbgarionuedp nanoanl-yssteastiswtiecraellyinsitghneifiscaamnet d(Tiraebclteion3).ofReefsfuelctst (fio.er. tthhee
ytisre there was no publication bias identified, no adjustment to pooled effect was negative), with variable statistical significance.
vin these values was made. In all of the subgroup analyses, the
U
e pooled effect sizes for LDL-C were negative, except for
gdirb when the five strata in which the baseline blood lipid level Discussion
ma wasoptimal werepooled.Forthesubgroupanalyses,statistic- In a meta-analysisthat included fiverandomised controltrials
C
yb ally significant pooled reductions in LDL-C were observed (and nine strata), Phung et al.(7) reported that almonds signifi-
e
niln when pooling those strata in which the almond dose was cantly reduce TC and have a strong trend towards reducing
o d ≥45g/d, the baseline LDL-C level was not optimal, the LDL-C (P=0·05). Although Phung et al.(7) also reported a
e
hsilb studydesignwascrossover,acontrolfood/dietwasprovided, near-significant reduction in HDL-C (P=0·08) and no effect
uP andthestudydurationwas<12weeks (Table3).Itshould be on TAG, in ouranalyses, which arebased on atotal of eight-
9
1.6 noted that when the control-adjusted changes in LDL-C for een publications and twenty-seven strata, the intake of
1
02 thethirteenstratainwhichthecontrolfood/dietwasnotpro- almonds was associated with significant reductions in TC,
.sn
j/7 vided were pooled, the reduction in LDL-C approached stat- LDL-C and TAG, and no effects on HDL-C.
101 istical significance (P=0·068). In a meta-analysis and dose–response of sixty-one con-
.01/g Almond consumption was not associated with any signifi- trolled intervention trials, which ranged in duration from 3
ro.io cant effect on fasting HDL-C, either in the overall analysis to 26 weeks, the consumption of nuts was associated with
d
//:sp 9
tth
https://doi.org/10.1017/jns.2016.19 Published online by Cambridge University Press
jo
u
r
n
a
ls
.c
a
Table3. Effectsofalmondsonbloodlipidlevels:resultsofmeta-analysesofrandomisedcontrolledtrials*† m
b
r
TC(mmol/l) LDL-C(mmol/l) HDL-C(mmol/l) TAG(mmol/l) TC:HDL-C LDL-C:HDL-C id
g
e
Allstrata n27 n25 n22 n25 n9 n10 .o
−0·153(−0·235,−0·070) −0·124(−0·196,−0·051) −0·017(−0·043,0·009) −0·067(−0·132,−0·002) −0·207(−0·362,−0·052) −0·089(−0·209,0·031) rg
P<0·001 P=0·001 P=0·207 P=0·042 P=0·009 P=0·145 /jn
Almonddose(g/d) ≥45 n17 n15 n13 n15 n6 n7 s
−0·212(−0·315,−0·108) −0·132(−0·209,−0·054) −0·020(−0·050,0·010) −0·071(−0·159,0·017) −0·173(−0·382,0·037) −0·065(−0·230,0·101)
P<0·001 P=0·001 P=0·188 P=0·114 P=0·106 P=0·445
<45 n10 n10 n9 n10 n3 n3
−0·039(−0·188,0·109) −0·060(−0·223,0·103) −0·008(−0·071,0·055) −0·064(−0·162,0·034) −0·260(−0·404,−0·116) −0·186(−0·279,−0·094)
P=0·605 P=0·470 P=0·808 P=0·199 P<0·001 P<0·001
Baselinelipidlevel Notoptimal n14 n20 n2;NA n7 NA NA
−0·271(−0·394,−0·148) −0·158(−0·238,−0·078) −0·189(−0·447,0·069)
P<0·001 P<0·001 P=0·151
Optimal n13 n5 n20 n16 NA NA
−0·044(−0·125,0·038) 0·100(−0·064,0·265) 0·003(−0·016,0·021) −0·034(−0·075,0·007)
P=0·294 P=0·232 P=0·773 P=0·100
Studydesign Crossover n10 n10 n10 n10 n6 n9
−0·182(−0·261,−0·103) −0·205(−0·316,−0·094) −0·017(−0·066,0·031) −0·017(−0·112,0·077) −0·147(−0·320,0·026) −0·138(−0·210,−0·066)
P<0·001 P<0·001 P=0·485 P=0·723 P=0·096 P<0·001
Parallel n17 n15 n12 n15 n3 n1;NA
−0·135(−0·268,−0·002) −0·048(−0·117,0·022) −0·014(−0·052,0·023) −0·111(−0·204,−0·017) −0·336(−0·693,0·021)
P=0·047 P=0·178 P=0·456 P=0·020 P=0·065
Controlfood/diet Provided n14 n12 n9 n12 n5 n9
−0·147(−0·241,−0·053) −0·155(−0·241,−0·069) 0·015(−0·008,0·038) −0·062(−0·132,0·008) −0·100(−0·266,0·067) −0·138(−0·210,−0·066)
P=0·002 P<0·001 P=0·189 P=0·081 P=0·241 P<0·001
Notprovided n13 n13 n13 n13 n3 n1;NA
−0·152(−0·293,−0·011) −0·093(−0·193,0·007) −0·028(−0·069,0·014) −0·085(−0·188,0·018) −0·386(−0·688,−0·084)
P=0·034 P=0·068 P=0·188 P=0·106 P=0·012
Duration(weeks) <12 n21 n19 n19 n19 n6 n9
−0·141(−0·210,−0·072) −0·151(−0·231,−0·071) −0·013(−0·048,0·022) −0·046(−0·100,0·009) −0·147(−0·320,0·026) −0·138(−0·210,−0·066)
P<0·001 P<0·001 P=0·468 P=0·101 P=0·096 P<0·001
≥12 n6 n6 n5 n6 n3 n1;NA
−0·169(−0·520,0·182) −0·031(−0·142,0·079) −0·027(−0·100,0·046) −0·155(−0·416,0·106) −0·336(−0·693,0·021)
P=0·346 P=0·576 P=0·467 P=0·245 P=0·065
TC,totalcholesterol;LDL-C,LDL-cholesterol;HDL-C,HDL-cholesterol;NA,notapplicable.
*Foreachmeta-analysis,theresultsindicate(inverticalorderfromtoptobottom):n(thenumberofstrata),thepooledeffect/pointestimate,the95%CI,andthestatisticalsignificanceofthepointestimate.Subgroupanalyseswereconductedif
therewereatleastthreeormorestrata.
†Anassessmentofpublicationbiaswasconductedforeachlipidparameter,butonlyforthemeta-analysisthatincludedallstrata.PublicationbiaswasnotidentifiedforTC,LDL-C,HDL-CorTAG.FortheratioofTC:HDL-C,twostudieswerefound
tobemissingtotherightofthepooledeffect,andwiththemissingstudiesimputed,thepooledeffectwas−0·120(95%CI−0·289,0·050).FortheratioofLDL-C:HDL-C,twostudieswerefoundtobemissingtotherightofthepooledeffect,and
withthemissingstudiesimputed,thepooledeffectwas−0·059(95%CI−0·175,0·056).
1
0