ORIGINAL RESEARCH
published: 23 May 2022
doi: 10.3389/ nu .2022.867690
F on ie s in Nu i ion | www. on ie sin.o g 1May 2022 | Volume 9 | A icle 867690
Edi ed by:
Daniel Moo e,
Uni e si y o To on o, Canada
Re iewed by:
S e an Kabisch,
Cha i é Uni e si ä smedizin
Be lin, Ge many
Hide aka Hamasaki,
Hamasaki Clinic, Japan
Jona han Pe e Li le,
Uni e si y o B i ish Columbia, Canada
*Co espondence:
Lukas Cip yan
[email p o ec ed]
o cid.o g/0000-0002-2403-8797
Special y sec ion:
This a icle was submi ed o
Spo and Exe cise Nu i ion,
a sec ion o he jou nal
F on ie s in Nu i ion
Recei ed: 01 Feb ua y 2022
Accep ed: 29 Ap il 2022
Published: 23 May 2022
Ci a ion:
Cip yan L, Li schmanno a M,
Ma e one PB, Plews DJ, Dos al T,
Ho mann P and Lau sen PB (2022)
Ve y Low-Ca bohyd a e High-Fa Die
Imp o es Risk Ma ke s o
Ca diome abolic Heal h Mo e Than
Exe cise in Men and Women Wi h
O e a Cons i u ion: Seconda y
Analysis o a Randomized Con olled
Clinical T ial. F on . Nu . 9:867690.
doi: 10.3389/ nu .2022.867690
Ve y Low-Ca bohyd a e High-Fa
Die Imp o es Risk Ma ke s o
Ca diome abolic Heal h Mo e Than
Exe cise in Men and Women Wi h
O e a Cons i u ion: Seconda y
Analysis o a Randomized Con olled
Clinical T ial
Lukas Cip yan1*, Ma ina Li schmanno a2, Philip B. Ma e one3, Daniel J. Plews4,
Tomas Dos al1, Pe e Ho mann5and Paul B. Lau sen4
1Depa men o Human Mo emen S udies & Human Mo ion Diagnos ic Cen e, The Uni e si y o Os a a, Os a a, Czechia,
2Depa men o Applied Ma hema ics, VSB—Technical Uni e si y o Os a a, Os a a, Czechia, 3Independen Resea che ,
B a lebo o, VT, Uni ed S a es, 4Spo s Pe o mance Resea ch Ins i u e New Zealand (SPRINZ), Auckland Uni e si y o
Technology, Auckland, New Zealand, 5Ins i u e o Human Mo emen Science, Spo & Heal h, Exe cise Physiology, T aining &
T aining The apy Resea ch G oup, Uni e si y o G az, G az, Aus ia
Pu pose: This andomized con olled pa allel-g oup s udy examined he e ec s o a e y
low-ca bohyd a e high- a (VLCHF) die and high-in ensi y in e al aining (HIIT) p og am
o e 12-weeks on ca diome abolic isk ac o s in indi iduals wi h o e a cons i u ion.
Me hods: Nine y-one pa icipan s ou o 109 comple ed he s udy. The pa icipan s
we e andomly alloca ed o he HIIT (N=22), VLCHF (N=25), VLCHF+HIIT
(N=25), o con ol (N=19) g oups o 12 weeks. Fas ing plasma samples we e
collec ed be o e he in e en ion and a e 4 and 12 weeks. The analyzed ou comes
included comple e blood coun , glucose, insulin, glyca ed hemoglobin, iglyce ides
(TG), choles e ol, high- and low-densi y lipop o ein (HDL-C and LDL-C), lipop o ein(a),
adiponec in (Adpn), lep in (Lep), umo nec osis ac o α(TNF-α), o he in e leukins
(hs-IL-6, IL-1β, and IL-10), and IL-1RA. The homeos asis model assessmen o insulin
esis ance (HOMA-IR), Adpn/Lep a io, TG/HDL-C a io, and TyG index we e calcula ed
and analyzed. Blood p essu e was measu ed be o e he in e en ion, a e 4, 8, and 12
weeks (ClinicalT ials.go : NCT03934476).
Resul s: Absolu e changes in HOMA-IR, Adpn/Lep a io, LDL-C, and dias olic blood
p essu e a e 12 weeks di e ed by s udy g oups (p<0.05). The mos p onounced
changes we e e ealed in he VLCHF (1M [95% CI]; HOMA-IR: −0.75 [−1.13; −0.55];
Adpn/Lep: 9.34 [6.33; 37.39]; LDL-C: 0.06 [−0.12; 0.50] mmol/l) and VLCHF+HIIT
(HOMA-IR: −0.44 [−1.14; 0.12]; Adpn/Lep: 4.26 [2.24; 13.16]; LDL-C: 0.25 [−0.04;
0.50] mmol/l) g oups.
Cip yan e al. VLCHF Die and HIIT
Conclusions: A 12-week VLCHF die in e en ion in indi iduals wi h o e a cons i u ion
is e ec i e o a o able changes in HOMA-IR (compa ed o HIIT), Adpn/Lep a io, and
dias olic blood p essu e. HIIT, o HIIT combined wi h he VLCHF die , had no addi ional
bene i s o he analyzed a iables. No ad e se side e ec s we e obse ed.
Keywo ds: HOMA-IR, adiponec in, lep in, TG/HDL-C, TyG index, low ca bohyd a e die , exe cise, o e a
INTRODUCTION
Physical ac i i y le els in he Wes e n popula ion ha e no
educed and changed li le, despi e d ama ic inc eases in he
o e a pandemic o he pas 30-plus yea s. Mo e han hal o
US adul s mee he ede al 2008 Physical Ac i i y Guidelines o
Ame icans and egula exe cise using ei he ae obic o muscle-
s eng hening exe cise, inc easing om 44% in 1998 o almos
52% in 2014 (1). Those mee ing hese guidelines o ae obic
ac i i y and muscle-s eng hening exe cise also inc eased om
abou 14% in 1998 o 21% in 2014. Howe e , a es o adul s
wi h o e weigh o obesi y ose o almos 71% du ing a simila
pe iod, e lec ing he o e a p e alence inc ease om 75% o
o e 90% (2).
Ca dio ascula and me abolic (ca diome abolic) isk ac o s
can con ibu e signi ican ly o inc eased mo bidi y and mo ali y,
educed quali y o li e, and highe heal hca e cos s. Th ee o he
majo isk ac o s include excess body a , low-g ade sys emic
ch onic in lamma ion, and insulin esis ance (IR). O e a is
de ined as excess body a ha impai s heal h (3). De e mina ion
o o e weigh and obese classi ica ions a e adi ionally based on
measu es o body mass index (BMI). This is no a di ec measu e
o body a and can misclassi y up o 50 % o mo e pa ien s wi h
bo h inc eased body a and i s associa ed disease isk ac o s.
The e o e, body a needs o be conside ed di ec ly o assess a
high- isk body composi ion (4).
O e a , and i s downs eam IR and ch onic in lamma ion,
which can main ain a die -induced iscous cycle, can also
lead o a wide ange o ca diome abolic heal h p oblems
such as me abolic synd ome, a he oscle osis, hype ension,
dyslipidemia, and ad anced ch onic condi ions such as Type 2
diabe es, ca dio ascula diseases, cance s, and neu odegene a i e
diseases (2). In addi ion, as ing iglyce ides and high-
densi y lipop o ein choles e ol, pa icula ly he a io o hese
wo measu es (TG/HDL-C), a e also conside ed a signi ican
ca diome abolic isk ac o as he a io e lec s IR (5). The
TG/HDL-C a io may be a be e clinical sc eening index han
he homeos asis model assessmen o IR (HOMA-IR) due o
accessibili y, ep oducibili y, and cos , and is al eady a commonly
used measu e in clinical p ac ice.
Exe cise and die a e wo commonly used modi iable li es yle
ac o s ha can help educe ca diome abolic isk ac o s o
in luence mo bidi y and mo ali y, imp o e quali y o li e,
and educe heal hca e cos s. While he impo ance o physical
ac i i y o inc eased i ness is undeniable, exe cise alone
may no necessa ily educe excess body a . In a p e ious
andomized con olled clinical ial we showed ha a e y low-
ca bohyd a e high- a (VLCHF) die alone educed excess body
a in indi iduals wi h o e a cons i u ion mo e han high-
in ensi y in e al aining (HIIT) alone (6). Low ca bohyd a e
die s a e e ec i e in emission o diabe es (7) and imp o e
insulin sensi i i y as measu ed by HOMA-IR (8–10). Simila ly,
insulin sensi i i y is also ela ed o he deg ee o physical
ac i i y. Exe cise has been shown o amelio a e insulin ac ion
in insulin- esis an indi iduals (11) by imp o emen o he
pa hophysiologic pa hways in ol ed in insulin esis ance (12).
The bene icial exe cise e ec on insulin sensi i i y occu s a e
se e al weeks (13) o e en a e a single bou o exe cise in
adul s wi h obesi y (14). Mo eo e , i seems ha HIIT induces
simila acu e imp o emen s in pe iphe al insulin sensi i i y as
mode a e-in ensi y con inuous aining (15).
A VLCHF die was p e iously shown o inc ease
adiponec in/lep in a io e lec ing educed sys emic low-
g ade in lamma ion in heal hy young indi iduals (16). In
con as , he same die was ound o be an e ec i e s a egy
o educing excess body a in men and women wi h o e a
cons i u ion (6). Sys ema ic e iews and me a-analyses show
also bene icial e ec s o low ca bohyd a e die s combined wi h
exe cise on body composi ion, iglyce ides, and ae obic capaci y
in adul s wi h obesi y (17,18). Consuming a high a die ,
especially one wi h high sa u a ed a y acids (SFA), is hough
o impai key aspec s o ca diome abolic heal h. This is despi e
no e idence-based associa ions be ween high in ake o SFA and
isk o a he oscle o ic p og ession (19). The e o e, he pu pose
o his andomized con olled pa allel-g oup s udy is o examine
he e ec s o a 12-week VLCHF die and high-in ensi y in e al
aining (HIIT) p og am on ca diome abolic isk ac o s in men
and women wi h o e a cons i u ion aged 20–59 yea s.
METHODS
Pa en S udy
I was a andomized, con olled, ou -a m, pa allel exe cise
and/o die a y in e en ion s udy (ClinicalT ials.go :
NCT03934476), wi h he p ima y aim o examining he VLCHF
and HIIT e ec on body composi ion and ca dio espi a o y
i ness le el (6). The me hod o andom assignmen is p esen ed
in Supplemen al Ma e ial. The e we e 91 pa icipan s alloca ed
o he ou s udy g oups and hese comple ed a 12-week
expe imen al pe iod (Figu e 1). Pa icipan s we e andomly
alloca ed o ou s udy g oups: 1) high-in ensi y in e al aining
(HIIT) and habi ual die , 2) e y low-ca bohyd a e, high- a
die (VLCHF) and habi ual physical ac i i y (no egula exe cise
aining), 3) VLCHF die and HIIT, and 4) Con ol (habi ual
die and physical ac i i y, no egula exe cise aining). Dual-
ene gy X- ay abso p iome y (DXA) and g aded exe cise es o
F on ie s in Nu i ion | www. on ie sin.o g 2May 2022 | Volume 9 | A icle 867690
Cip yan e al. VLCHF Die and HIIT
FIGURE 1 | Flow cha (6). Bold alue indica es numbe o pa icipan s.
F on ie s in Nu i ion | www. on ie sin.o g 3May 2022 | Volume 9 | A icle 867690
Cip yan e al. VLCHF Die and HIIT
oli ional exhaus ion we e used o he body composi ion and
ca dio espi a o y i ness (CRF) assessmen s, espec i ely.
To ob ain measu es o no in e en ion, a con ol g oup was
u ilized. Pa icipan s in he con ol g oup we e ad ised no o
change hei habi ual die and physical ac i i y egime. The e o e,
no die ad ice was p o ided.
Resul s o he p ima y ou come we e p e iously epo ed,
ha a VLCHF die , ei he in isola ion o in combina ion wi h
HIIT, caused a signi ican educ ion in isce al adipose issue
(VAT) mass and body composi ion a iables. HIIT alone did
no induce such e ec s on body composi ion, bu imp o ed
exe cise capaci y (6). We u ilized he in as uc u e o his ial
o conduc a p eplanned ancilla y s udy ocused on clinically
ele an isk ac o s o ca diome abolic heal h. We analyzed
blood samples ollowing a 3-h as be o e he expe imen al pe iod
(T0) and a e 4, and 12 weeks (T1and T3). Blood p essu e was
analyzed also a e 8 weeks (T2). The pa icipan se used o he
p ima y analysis was iden ical wi h he pa icipan se p esen ed
in his s udy.
Pa icipan s
We en olled adul s aged 20–59 yea s wi h BMI 25.00–40.00
kg/m2, who we e no engaged in any egula exe cise. Pa icipan s
wi h known ch onic diseases we e excluded. Addi ional eligibili y
c i e ia a e lis ed in Supplemen a y Table 1. The pa icipan s
had no p e ious expe ience wi h he VLCHF die o HIIT. The
ec ui men de ails and d opou s du ing he s udy a e shown
in Figu e 1. W i en in o med consen was ob ained om all
s udy pa icipan s. The s udy design was app o ed by he Os a a
uni e si y E hics Commi ee (n . 1/2018).
High-In ensi y In e al T aining (HIIT)
P io o he in e en ion, he pa icipan s we e p o ided wi h
de ailed ins uc ions on he HIIT p og am. This was o bo h
he HIIT and VLCHF+HIIT g oups and done bo h in e bally
and w i en o m. The pa icipan s we e ins uc ed o comple e 3
HIIT sessions pe week whe e one HIIT session was comple ed
du ing weeks 4, 8, and 12 when he pa icipan s isi ed he
labo a o y and wo we e home-based and sel -pe o med. Each
HIIT session had a wa m up and cool down pe iod o 5-min o
slow walking. HIIT consis ed o a 3 min in e al o high-in ensi y
walking (Bo g’s scale RPE 18–19) ollowed by a 3 min in e al o
low-in ensi y walking (RPE 9–11). Pa icipan s pe o med 4, 6,
and 8 high-in ensi y in e als in he i s , second, and hi d 4-
week pe iod, espec i ely. The e o e, du a ion a high-in ensi y
was 12, 18 and 24 min and o al session ime inc eased om
31 o 43 min and 55 min du ing each 4-weeks, espec i ely.
T aining in ensi y was measu ed wi h a hea a e moni o (Pola
M430; Pola Elec o, Oy, Finland). These da a we e subsequen ly
uploaded o Pola Flow (Pola Elec o, Finland) and analyzed
egula ly o ack compliance. Pa icipan s we e ins uc ed o
eco d all addi ional aining sessions o any ype in addi ion o
he s udy p o ocol.
Die a y In e en ion
Bo h he HIIT and con ol g oups we e asked o main ain hei
habi ual die a y in ake wi hou es ic ion. The VLCHF die was
de ined as allowing no mo e han 50 g o ca bohyd a es (CHO)
pe day (20). Nei he die included a speci ic calo ie o ene gy
goal. Howe e , pa icipan s in he VLCHF g oup we e ad ised
o compensa e o he o al ene gy dec ease caused by CHO
in ake es ic ion by inc easing hei na u al non- ans- a in ake
(e.g., c eam, bu e , oli e, and coconu oil). A a ge p o ein
in ake o 1.5 g/kg lean body mass was ecommended. Con a y
o he s ic CHO es ic ion, pa icipan s we e asked o keep o
a ge s. The use o all swee ened and g ain-based p oduc s had
o be minimized. The ecommended ood included whole ood
sou ces, such as mea s, ege ables, non-swee ened p oduc s, ull-
a dai y i ems, nu s, and seeds. A die i ian p o ided de ailed
die a y ad ice be o e and du ing he s udy (on eques o a
leas once a mon h). In addi ion, a handbook was p o ided
o pa icipan s con aining ood lis s, guidelines o es ima ing
mac onu ien amoun s, and sample ecipes. To eco d all oods
and quan i ies consumed an app was used in all s udy g oups
(www.kalo icke abulky.cz). This commenced se en days be o e
he s a o he in e en ion. Alcoholic be e ages we e es ic ed
du ing he in e en ion pe iod, and die a y supplemen s we e
no pe mi ed 1 mon h be o e and du ing he in e en ion
pe iod. Ca eina ed be e ages we e es ic ed only be o e he
labo a o y sessions.
An h opome ic Analysis
The esul s o he an h opome ic analysis ha e p e iously been
epo ed (6). In summa y, he o al body mass and isce al
adipose issue (VAT) mass signi ican ly dec eased in he VLCHF
(by median [IQR]: −6.9 ([−8.4; −5.6]) % and −23.2 [−26.5;
−14.7] %, espec i ely) and VLCHF+HIIT (by −9.0 [–10.9;
−7.9] % and −17.6 [−23.8; −10.8] %, espec i ely) g oups
a e 12 weeks despi e no signi ican changes in he HIIT and
Con ol g oups.
Labo a o y Me hods
Fas ing blood samples we e collec ed om he an ecubi al ein.
Whole blood samples wi h EDTA as an an icoagulan we e used
immedia ely o blood coun and HbA1c de e mina ion. Se um
collec ion ubes we e allowed o clo o 30 min and subsequen ly
cen i uged a 2 500 g o 10 min o sepa a e he se um. Blood
se um was di ided in o h ee 1-ml aliquo s, which we e ozen
a −80◦C un il analysis. The S-Mono e e R
sys em (Sa s ed ,
Nümb ech , Ge many) was used o blood sample collec ion.
Blood coun pa ame e s we e measu ed using a UniCel R

DxHTM 800 hema ology analyze (Beckman Coul e , Inc., B ea,
CA, USA). Glyca ed hemoglobin (HbA1c) was measu ed using
a D-10TM Bio-Rad de ice (Bio-Rad Labo a o ies, Inc., He cules,
CA, USA). Glucose, iglyce ide (TG), o al choles e ol, and
high- and low-densi y lipop o ein choles e ol (HDL-C and LDL-
C, espec i ely) concen a ions we e measu ed using an AU
5820 de ice (Beckman Coul e , Inc., B ea, CA, USA). Se um
le els o lep in, adiponec in, TNF-α, IL−1RA, IL−1β, and IL-
10 we e de e mined by mul iplex echnology using a Bio-
Plex MAGPIX sys em (Bio-Rad Labo a o ies, Redmond, WA,
USA). Hs-IL-6 concen a ions we e measu ed using a Human
IL-6 Quan ikine ELISA ki (R&D Sys ems, Minneapolis, MN,
USA) on a DSX de ice (Dynex Technologies, Chan illy, VA,
F on ie s in Nu i ion | www. on ie sin.o g 4May 2022 | Volume 9 | A icle 867690
Cip yan e al. VLCHF Die and HIIT
USA). Insulin concen a ion was measu ed using a UniCel
DxI 800 analyze (Beckman Coul e , Inc., B ea, CA, USA).
Lipop o ein(a) [Lp(a)] concen a ion was measu ed using a BN
P oSpec sys em (BN P oSpec, Siemens Heal hca e Diagnos ics
P oduc GmbH, Ge many).
The in a-assay coe icien s o a ia ion o biochemical and
blood coun pa ame e s we e <5%. Lep in, adiponec in, TNF-α,
IL−1RA, IL−1β, IL−10, and hs-IL−6 we e de e mined wi h an
in e -assay coe icien lowe han 10%.
T iglyce ide-glucose (TyG) index was calcula ed by applying
he ollowing equa ion (21):
TyGindex =ln  as ing se um TG × as ing plasma glucose
2(1)
Homeos a ic model assessmen o insulin esis ance (HOMA-IR)
was calcula ed using he o mula (22):
HOMA −IR =plasma glucose ×se um insulin
22.5 (2)
A capilla y blood sample was d awn om a inge o
measu e β-hyd oxybu y a e (βHB) (F eeS yle Op ium Neo,
Oxon, Uni ed Kingdom). All pa icipan s sel -analyzed i wice
a week (e e y Monday and Thu sday) in a as ing s a e in he
mo ning o moni o esponses o he VLCHF die and o con ol
o adhe ence.
Blood P essu e
Sys olic and dias olic blood p essu e (BP) was au oma ically
measu ed h ee imes wi h 1–2 min apa a e he pa icipan
had been si ing o ≥10 min in a quie oom by applying a
s anda d de ice (Nissei DM 3000, Nihon Seimi su Sokki Co.,
Japan). This p ocedu e is in line wi h he ecommenda ions o he
Ame ican College o Ca diology, Ame ican Hea Associa ion,
and Eu opean Socie y o Hype ension (23,24). Pa icipan s we e
ins uc ed o a oid ca eina ed be e ages o a leas 60 min be o e
he blood p essu e measu emen s.
S a is ical Analyses
The ca ego ical a iable (sex) was desc ibed by equency
a io, and nume ical a iables we e desc ibed by median and
in e qua ile ange (IQR) a each ime poin . Subsequen ly, he
absolu e changes o he moni o ed a iables a ime T1, T2,
and T3wi h espec o he baseline (T0) we e analyzed. The
absolu e changes we e es ed o no mal dis ibu ion using he
Shapi o-Wilk es . In some cases, signi ican de ia ions om
no mali y we e de ec ed such ha non-pa ame ic me hods o
da a desc ip ion (median and in e qua ile ange) and s a is ical
in e ence we e used. Signi icance o change was es ed by 95%
con idence in e al (CI) o median and wo- ailed Wilcoxon
signed- ank es o each a iable, each g oup, and each ime.
The e ec size (ES) o he obse ed changes was speci ied by he
Wilcoxon e ec size ( ), including i s 95% con idence in e al.
Th eshold alues o ES we e 0.10 o <0.30 (small), 0.30 o <
0.50 (medium), ≥0.50 (la ge).
Finally, he absolu e changes in he gi en a iables o
he HIIT, VLCHF, HIIT+VLCHF, and Con ol g oups we e
compa ed using he K uskal-Wallis es a each ime poin .
Dunn’s es was used o analyze speci ic sample pai s o
s ochas ic dominance. Dunn’s es mul iple compa ison p- alues
was adjus ed wi h he Benjamini-Hochbe g me hod. The e ec
size o he obse ed di e ences was assessed using he e a squa ed
based on he H-s a is ic, including i s 95% con idence in e al.
Th eshold alues o ES e a squa ed we e 0.01 o <0.08 (small),
0.08 o <0.26 (medium), ≥0.26 (la ge) (25).
An a p io i powe analysis using GPOWER (26) wi h powe
se a 0.80 and signi icance le el se a 0.05 was calcula ed
e ospec i ely. The powe analysis indica ed ha a o al sample
o 76 people would be needed o de ec la ge e ec s ( =0.40)
o his s udy wi h 4 g oups. A o al sample o 180 people would
be needed o de ec medium e ec s ( =0.25) (27). Thus, he
sample size was su icien o e eal ha a la ge e ec could no be
in e p e ed as non-signi ican .
In all cases, s a is ical signi icance was se a p<0.05.
S a is ical analyses we e pe o med using R Co e Team (28).
RESULTS
Pa icipan s
The low cha o pa icipan s h ough he ial, as well as
he easons o d opou s, a e depic ed in Figu e 1. Pa icipan
cha ac e is ics a baseline a e lis ed in Table 1.
Die
To al ene gy in ake dec eased (p<0.05) in he HIIT (median
[95% CI]: −6.1 [−0.2; −13.4] %), VLCHF (−19.7 [−12.5;
−25.2] %), and VLCHF+HIIT (−25.8 [−20.5; −28.0] %) g oups.
Ca bohyd a e in ake dec eased (p<0.05) by −81.8 [−79.1;
−82.9] % and −82.8 [−80.4; −85.7] % in he VLCHF and
VLCHF+HIIT g oups, espec i ely. Fa in ake inc eased by
44.6 [36.1; 61.7] % and 34.8 [24.6; 47.3] % in he VLCHF
and VLCHF+HIIT g oups, espec i ely. P o ein in ake did no
signi ican ly change in any o he s udy g oups. To al ene gy,
p o ein, and ca bohyd a e in ake did no signi ican ly change in
he con ol g oup, whe eas a in ake dec eased (p=0.023, −6.0
[−1.0; −17.5] %) (Supplemen a y Table 2).
High-In ensi y In e al T aining
The e we e subs an ial be ween-g oup di e ences in he
aining cha ac e is ics. To al aining ime in he HIIT
and VLCHF+HIIT g oups (median 1424 and 1452 min,
espec i ely) was subs an ially highe han hose wi hou he
HIIT in e en ion (VLCHF−124 min, Con ol−105 min). A
de ailed aining session analysis has al eady been published in
he pa en s udy (6).
Biochemical Analysis
The e we e no signi ican be ween-g oup di e ences in all he
biochemical a iables a baseline.
Absolu e changes in HOMA-IR a e 12 weeks di e ed by
s udy g oup (p=0.013; ES 95% CI: small o la ge). Howe e ,
no in e en ion g oup signi ican ly di e ed om he Con ol
F on ie s in Nu i ion | www. on ie sin.o g 5May 2022 | Volume 9 | A icle 867690

Cip yan e al. VLCHF Die and HIIT
TABLE 1 | Baseline cha ac e is ics o s udy pa icipan s (all Caucasian).
HIIT (N=22) VLCHF (N=25) VLCHF+HIIT (N=25) Con ol (N=19) Be ween-g oup di e ences
M(IQR) M(IQR) M(IQR) M(IQR) p- alue
Male:Female 6:16 8:17 7:18 6:13 0.648
Age (yea ) 46 (38.8; 53.3) 43 (35.0; 51.5) 43 (32,51) 40 (31; 53) 0.722
Heigh (cm) 167.7 (160.6; 175.0) 170.1 (164.7; 177.9) 169.8 (160.9; 179.7) 171.9 (162.8; 177.6) 0.336
BMI 28.7 (26.9; 30.9) 31.3 (27.7; 33.0) 31.0 (27.2; 35.1) 28.7 (26.7; 32.9) 0.690
WH R 0.60 (0.54; 0.62) 0.62 (0.56; 0.65) 0.62 (0.54; 0.68) 0.60 (0.54; 0.63) 0.447
Sys olic BP (mmHg) 127 (114; 137) 132 (119; 144) 128 (121; 138) 124 (121; 134) 0.105
Dias olic BP (mmHg) 77 (72; 87) 88 (77; 93) 85 (80; 91) 82 (73; 88) 0.919
Hemoglobin (g/l) 137 (135; 145) 140 (132; 150) 140 (130; 148) 137 (130; 150) 0.934
Hema oc i (%) 0.406 (0.396; 0.421) 0.411 (0.389; 0.441) 0.41 (0.378; 0.433) 0.405 (0.381; 0.435) 0.969
E y h ocy es (1012/l) 4.57 (4.4; 4.81) 4.72 (4.29; 5.07) 4.62 (4.36; 4.91) 4.56 (4.44; 4.88) 0.789
Th ombocy es (109/l) 244 (217; 294) 241 (210; 274) 234 (200; 267) 239 (202; 280) 0.623
Leukocy es (109/l) 6.5 (5.9; 7.6) 6.5 (5.9; 7.9) 6.4 (5.3; 7) 6.7 (5.4; 7.6) 0.466
HbA1c (mmol/mol) 34 (31.5; 38.5) 34 (32; 37.8) 36 (34; 38) 34 (30; 37) 0.300
Glucose (mmol/l) 5.17 (4.94; 5.4) 5.21 (5.01; 5.59) 5.12 (4.42; 5.34) 4.91 (4.65; 5.43) 0.578
T iglyce ides (mmol/l) 1.18 (0.86; 1.81) 1.28 (0.86; 2.23) 0.96 (0.77; 2.18) 1.68 (0.79; 2.32) 0.410
Choles e ol (mmol/l) 5.21 (4.51; 5.93) 5.55 (5.15; 6.14) 5.11 (4.39; 6.05) 5.47 (4.91; 6.16) 0.764
HDL-C (mmol/l) 1.36 (1.17; 1.55) 1.29 (1.04; 1.53) 1.18 (1.07; 1.47) 1.32 (1.02; 1.57) 0.210
LDL-C (mmol/l) 3.3 (2.61; 3.76) 3.49 (3.03; 3.9) 3.16 (2.49; 3.59) 3.47 (2.95; 3.79) 0.386
Insulin (mU/l) 7.7 (5.2; 10.2) 9.6 (6.2; 14.6) 7.3 (5.3; 12.2) 8.2 (7.3; 14.1) 0.436
Lep in (ng/l) 5.7 (2.8; 7.4) 6.9 (3.4; 12.9) 6.2 (4.6; 12.6) 6.2 (3; 12.3) 0.579
Adiponec in (mg/l) 27 (21; 39) 41 (22; 59) 42 (16; 58) 27 (15; 59) 0.735
TG/HDL-C (–) 0.77 (0.56; 1.37) 1.07 (0.6; 2.05) 0.74 (0.55; 1.9) 1.15 (0.47; 2.32) 0.585
TyG index (–) 3.06 (2.25; 4.67) 3.24 (2.32; 5.9) 2.63 (1.86; 5.68) 4 (2.15; 5.4) 0.398
HOMA-IR (–) 1.78 (1.13; 2.5) 2.03 (1.44; 3.66) 1.51 (1.13; 3.11) 1.92 (1.51; 3.51) 0.726
Adpn/Lep (–) 5.66 (3.39; 9.98) 5.06 (3.29; 9.53) 4.19 (2.53; 9.61) 7.16 (2.24; 22.11) 0.648
Legend: BMI, body mass index; WH R, wais - o-heigh a io; BP, blood p essu e; HbA1c, glyca ed haemoglobin; TG, iglyce ides; HDL-C/LDL-C, high/low densi y lipop o ein; TyG
index, iglyce ide-glucose index; HOMA-IR, homeos a ic model assessmen o insulin esis ance; Adpn/Lep, adiponec in-lep in index. Da a a e he median (M) wi h in e qua ile ange
(IQR). K uskal-Wallis es o he be ween-g oup di e ences.
g oup. The be ween-g oup signi ican di e ences we e caused
by he di e ences be ween HIIT and VLCHF g oups. The
mos subs an ial HOMA-IR dec ease was in he VLCHF g oup
(median [IQR]: −35.7 [– 20.2; −44.0] %). This dec ease in
he VLCHF g oup was caused by changes o bo h HOMA-
IR componen s insulin and glucose. Unlike glucose changes,
absolu e insulin changes di e ed by s udy g oup (p=0.023, ES
95% CI: small o la ge) (Table 2 and Figu e 2).
Absolu e changes in he adiponec in/lep in (Adpn/Lep) a io
a e 12 weeks di e ed by s udy g oup (p <0.001; ES 95% CI:
la ge) wi h he di e ences be ween he HIIT and Con ol g oups
s. he VLCHF and VLCHF+HIIT g oups. The Adpn/Lep a io
inc eased in he VLCHF g oup by 120.4 [88.7; 287.1] % and
VLCHF+HIIT g oup by 158.9 [49.5; 540.2] %. These Adpn/Lep
inc eases in he VLCHF and VLCHF+HIIT g oups we e caused
by bo h he lep in dec eases (p<0.001; ES 95% CI: la ge) and
adiponec in inc eases (p=0.054; ES 95% CI: small o la ge)
(Table 2 and Figu e 2).
The TyG index and TG/HDL-C a io signi ican ly dec eased
in he VLCHF (-0.74 [-2.11; −0.27] and −0.13 [−0.40; −0.04],
espec i ely) and VLCHF+HIIT (−0.68 [−1.89; 0.00] and −0.18
[−0.65; −0.03]) g oup (Table 2 and Figu e 2).
Absolu e changes in LDL-C a e 12 weeks di e ed by s udy
g oup (p=0.003; ES 95% CI: small o la ge) A pos hoc analysis
e ealed di e ences be ween he bo h die g oups (VLCHF,
VLCHF+HIIT) and HIIT and Con ol g oups. LDL-C non-
signi ican ly changed in he VLCHF g oup by 1.6 [−7.8; 20.0]
% and in he VLCHF+HIIT g oup by 7.4 [−6.1; 20.3] %).
Howe e , LDL-C dec eased in he HIIT (p=0.012; −8.7
[−12.7; 0.3] %) and Con ol (p=0.016; −10.1 [−18.3; −1.0]
%) g oups (Table 2). The comple e da ase is epo ed in he
Supplemen a y Ma e ial (Supplemen a y Tables 3–5).
Lp(a), TNF-α, hs-IL−6, IL−1RA, IL−1β, and IL-10 emained
mos ly unde a de ec ion le el o he assay. The e o e, no u he
s a is ical analyses we e conduc ed (Supplemen a y Table 6).
The e we e subs an ial inc eases in β-hyd oxybu y a e
concen a ion (βHB) in he VLCHF and VLCHF+HIIT g oups.
The highes βHB concen a ions we e achie ed a e 2 weeks
o VLCHF die in e en ion. βHB concen a ions in he HIIT
and con ol g oups emained wi hin he ange be ween 0.0 o
F on ie s in Nu i ion | www. on ie sin.o g 6May 2022 | Volume 9 | A icle 867690
Cip yan e al. VLCHF Die and HIIT
FIGURE 2 | Absolu e changes in HOMA-IR, Adpn/Lep a io, TG/HDL-C a io, and TyG index a e 4 (T1) and 12 (T3) weeks.
F on ie s in Nu i ion | www. on ie sin.o g 7May 2022 | Volume 9 | A icle 867690
Cip yan e al. VLCHF Die and HIIT
TABLE 2 | Biochemical a iables di e ences a e 12 weeks.
HIIT VLCHF VLCHF+HIIT Con ol Be ween–g oup di . (p– alue)
1M(95% CI) 1M(95% CI) 1M(95% CI) 1M(95% CI)
Hemoglobin (g/l) −1.5 (−4.5; 1) −1 (−3.5; 0)* −4 (−6; −1.5)* −2 (−4; 1) 0.419
Hema oc i (%) −0.01 (−0.01; 0.00)* 0.00 (−0.01; 0.00) −0.01 (−0.01; 0.00)* −0.01 (−0.01; 0.00) 0.705
E y h ocy es (1012/l) −0.08 (−0.16; 0.02) −0.03 (−0.11; 0.02) −0.12 (−0.20; −0.06)* −0.07 (−0.17; 0.02) 0.343
Th ombocy es (109/l) 2.5 (−9.5; 11.5) −8.0 (−15.0; 6.5) −25.0 (−29.0; −8.0)* 3.0 (−13.5; 15.0) 0.030a
Leukocy es (109/l) −0.40 (−0.65; 0.25) 0.00 (−0.65; 0.40) −0.30 (−1.10; 0.15) −0.10 (−0.85; 0.55) 0.760
HbA1c (mmol/mol) 0.0 (−2.0; 2.0) −2.0 (−3.0; 0.0) −2.0 (−4.0; 0.0)* 1.0 (−0.5; 3.0) 0.081
Glucose (mmol/l) −0.02 (−0.38; 0.27) −0.31 (−0.54; −0.05)* 0.01 (−0.50; 0.33) −0.21 (−0.58; −0.04)* 0.316
T iglyce ides (mmol/l) 0.03 (−0.33; 0.32) −0.28 (−0.58; −0.10)* −0.20 (−0.70; −0.01)* 0.10 (−0.41; 0.30) 0.161
Choles e ol (mmol/l) −0.46 (−0.56; −0.05)* −0.01 (−0.28; 0.56) 0.29 (−0.22; 0.54) −0.21 (−0.82; 0.01) 0.063
HDL–C (mmol/l) −0.09 (−0.19; 0.03) 0.03 (−0.08; 0.07) 0.07 (−0.05; 0.20) −0.03 (−0.20; 0.04) 0.088
LDL–C (mmol/l) −0.33 (−0.42; −0.08)* 0.06 (−0.12; 0.50) 0.25 (−0.04; 0.50) −0.33 (−0.57; −0.06)* 0.003b
Insulin (mU/l) 0.86 (−1.61; 5.27) −2.84 (−3.91; −1.69)** −2.14 (−4.54; 0.37) 0.68 (−2.40; 2.32) 0.023c
Lep in (ng/l) −0.19 (−1.36; 0.71) −3.31 (−5.69; −2.59)** −2.40 (−4.42; −1.31)* 4.07 (2.86; 5.53)** <0.001d
Adiponec in (mg/l) −1.70 (−8.48; 2.96) 9.30 (1.88; 27.85)* 3.55 (−3.85; 6.41) 3.62 (−43.18; 13.94) 0.054
TG/HDL–C (–) 0.03 (−0.21; 0.29) −0.13 (−0.40; −0.04)* −0.18 (−0.65; −0.03)* 0.04 (−0.16; 0.31) 0.060
TyG index (–) −0.16 (−0.87; 10) −0.74 (−2.11; −0.27)* −0.68 (−1.89; 0.00)* −0.12 (−1.18; 0.56) 0.207
HOMA–IR (–) 0.19 (−0.44; 1.24) −0.75 (−1.13; −0.55)** −0.44 (−1.14; 0.12) −0.01 (−0.73; 0.38) 0.013e
Adpn/Lep (–) −0.08 (−1.46; 1.03) 9.34 (6.33; 37.39)** 4.26 (2.24; 13.16)* −3.24 (−21.27; −0.83)* <0.001b
Legend: HbA1c, glyca ed hemoglobin; TG, iglyce ides; HDL–C/LDL–C, high/low densi y lipop o ein; TyG index, iglyce ide–glucose index; HOMA–IR, homeos a ic model assessmen
o insulin esis ance; Adpn/Lep, adiponec in/lep in a io.
Da a a e he median di e ences (1M) be ween baseline minus 12–week measu es wi h 95% con idence in e als (CI). The comple e da ase is epo ed in he Supplemen a y Ma e ial.
Two– ailed Wilcoxon signed– ank es : *signi ican di e ences (p<0.05) o baseline s. 12–week; ** signi ican di e ences (p<0.001) o baseline s. 12–week.
K uskal–Wallis es o he be ween–g oup di e ences. Pos –hoc analysis (homogenous subg oups): a– (VLCHF and VLCHF+HIIT), (HIIT, VLCHF and Con ol); b– (HIIT and Con ol),
(VLCHF and VLCHF+HIIT); c– (VLCHF and VLCHF+HIIT), (HIIT, VLCHF+HIIT and Con ol); d– Con ol, HIIT, (VLCHF and VLCHF+HIIT); e– (VLCHF, VLCHF+HIIT and Con ol), (HIIT,
VLCHF+HIIT and Con ol).
0.3 mmol/l o he whole 12-week in e en ion. A de ailed βHB
analysis has al eady been p esen ed in he pa en s udy (6).
Blood P essu e
Sys olic BP dec eased (p <0.05) in all h ee in e en ion g oups
a e 8 weeks, when compa ed o he baseline le el, and emained
signi ican ly dec eased a e 12 weeks in he VLCHF+HIIT
g oup. Howe e , sys olic BP did no signi ican ly di e by s udy
g oup a any ime poin . The be ween-g oup di e ences we e
shown in dias olic BP a e 12 weeks (p=0.049, ES 95% CI: small
o medium), wi h he mos p onounced dec eases in he VLCHF
(1M[95% CI]: −4.0 [−6.8; −0.3] mmHg; ES 95% CI: small o
la ge) and VLCHF+HIIT (−5.3 [−8.0 −3.3] mmHg; ES 95% CI:
la ge) g oups (Table 3).
DISCUSSION
In his andomized con olled ial, we ound ha VLCHF die ,
when compa ed o HIIT, o e 12 weeks in indi iduals wi h
o e a cons i u ion had subs an ial bene i s o ch onic non-
communicable diseases isk ac o s HOMA-IR, Adpn/Lep a io
and dias olic BP beside he al eady p esen ed dec ease in body
mass and VAT (6). Howe e , we did no ind signi ican changes
o HOMA-IR om he Con ol g oup. Adding HIIT o he
VLCHF die did no caused an ex a e ec on hese a iables. We
showed ha VLCHF die imp o ed insulin sensi i i y and skewed
he lep in and adiponec in le els owa d an i-in lamma o y
pheno ypes. Despi e signi ican inc eases in sa u a ed a in ake,
we ound no signi ican ele a ion o LDL-C in he VLCHF and
VLCHF+HIIT g oups.
VLCHF Die and βHB
Less han 50 g/day o CHO was equi ed o he VLCHF
g oups in his s udy (29). The aim o he VLCHF die
in e en ion was no o educe o al ene gy in ake. As such
he VLCHF and VLCHF+HIIT g oups we e encou aged o
compensa e o he CHO in ake es ic ion by inc easing a
in ake while main aining p o ein consump ion. Ne e heless, a
in ake was insu icien o keep he o al ene gy in ake unchanged
(Supplemen a y Table S2). This is indeed a common si ua ion
in eal-li e condi ions. ßHB measu es con i med adhe ence o
he die in bo h VLCHF and VLCHF+HIIT g oups as has been
shown al eady (13). No ably, o al ene gy in ake signi ican ly
dec eased in he HIIT g oup despi e no die modi ica ion. This
e ec was likely due o an inc eased in e es in a heal hy li es yle
when pa icipa ing in such a esea ch s udy.
HOMA-IR
We ound ha he HOMA-IR, a equen ly used index o
e alua e insulin esis ance, was subs an ially educed a e 12
weeks o he VLCHF die , when compa ed o HIIT. We showed
his imp o ed insulin sensi i i y e en i he pa icipan s wi h
F on ie s in Nu i ion | www. on ie sin.o g 8May 2022 | Volume 9 | A icle 867690
Cip yan e al. VLCHF Die and HIIT
TABLE 3 | Blood p essu e ou comes.
HIIT VLCHF VLCHF+HIIT Con ol Be ween–g oup di . (p– alue)
Sys olic
PRE 127 (114; 137) 132 (119; 144) 128 (121; 138) 124 (121; 134) –
4 weeks 121 (112; 134) 126 (118; 132)* 125 (115; 134) 124 (117; 131) 0.339
8 weeks 118 (112; 134)* 124 (114; 133)* 120 (114; 134)* 120 (112; 126) 0.990
12 weeks 124 (115; 140) 127 (117; 138) 125 (115; 130)* 120 (118; 133) 0.236
Dias olic
PRE 77 (72; 87) 88 (77; 93) 85 (80; 91) 82 (73; 88) –
4 weeks 79 (72; 87) 82 (79; 87) 81 (77; 90)* 79 (74; 86) 0.137
8 weeks 75 (69; 84)* 83 (79; 88) 80 (72; 86)** 77 (71; 83) 0.380
12 weeks 80 (72; 87) 83 (75; 89)* 81 (73; 86)** 79 (73; 82) 0.049a
Legend, *di e en om he baseline (PRE) a p <0.05.
Values a e shown as median (in e qua ile ange).
Two– ailed Wilcoxon signed– ank es , *signi ican di e ences (p<0.05) o baseline (PRE); ** signi ican di e ences (p<0.001) o baseline (PRE).
K uskal–Wallis es o he be ween–g oup di e ences. Pos –hoc analysis (homogenous subg oups): a– (HIIT,VLCHF and Con ol), (VLCHF, VLCHF+HIIT and Con ol).
o e a cons i u ion we e wi hou diabe es and wi hin he no mal
ange o HbA1c. Low ca bohyd a e die s p o ed o be mo e
e ec i e han highe ca bohyd a e (low a ) die s in imp o ing
as ing glucose and insulin and insulin sensi i i y as measu ed
by HOMA-IR in indi iduals wi h obesi y and insulin esis ance
(9) and pa ien s wi h obesi y and non-alcoholic a y li e disease
(10). No su p isingly, low ca bohyd a e die s a e associa ed wi h
a la ge (32 %) inc ease in emission o diabe es (7). Insulin
esis ance and excessi e body a a e conside ed among he mos
impo an causes o se e al ch onic me abolic and ca dio ascula
diseases. The cellula and physiological mechanisms a e complex
and in ol e adiposi y-induced al e a ions in βcell unc ion,
adipose issue biology, and mul i-o gan insulin esis ance. All
hese pe spec i es can be imp o ed wi h adequa e body mass loss
(30), which we also showed in his s udy (6) which may ha e
con ibu ed o he HOMA-IR educ ion.
Ano he su oga e measu e o he diagnosis o insulin
esis ance is he TyG index, which is independen ly and mo e
s ongly associa ed wi h a e ial s i ness in pa ien s wi h ype
2 diabe es han HOMA-IR (31). The TyG index signi ican ly
dec eased in bo h he VLCHF g oup by median 25.9 [IQR: −38.0;
1.2] % and VLCHF+HIIT g oup by median 23.1 [−54.7; 27.0]
%, when he 12-week ou comes we e compa ed o he baseline.
Howe e , hese a o able changes we e no su icien o p o e a
signi ican be ween-g oup di e ences (p=0.207). Ne e heless,
we can sugges ha a ca bohyd a e in ake es ic ion migh
be bene icial no only o he isce al adipose issue educ ion
as we showed in he pa en s udy (6), bu also o he
ea men o he impai ed insulin esis ance, as al eady shown
elsewhe e (29,32,33).
The uniqueness o his s udy, howe e , lay in including
HIIT, alone o in combina ion wi h VLCHF die , in o
conside a ion. An exe cise in e en ion p og am p o ed o be
e ec i e in he ea men o insulin esis ance in indi iduals
wi h o e weigh /obesi y (34), me abolic synd ome (35) o ype
2 diabe es, i.e., educes as ing insulin, HOMA-IR, as ing
blood suga , HbA1c, and body mass index (36). Howe e ,
we did no show any o addi ional e ec o HIIT on hese
a iables a e 12 weeks. This inconsis ency can be ela ed o
pa icipan s cha ac e is ics, s udy du a ion o o he design issues.
A die adjus men seems o be, he e o e, c ucial wi hin any
li es yle modi ica ion o body mass managemen and educing
heal h isk a iables, despi e we s ill conside physical ac i i y
and egula exe cise impo an , e.g., o he main enance o
imp o emen o he CRF le el (6). The e is a solid e idence ha
he CRF le el is in e sely associa ed wi h all-cause, CVD and
cance mo ali y. A dose- esponse analyses e en showed ha a
pe one-MET inc ease o he CRF le el was associa ed wi h 12
%, 13 %, and 7 % educed isk o all-cause, CVD and cance
mo ali y (37).
Adpn/Lep Ra io
We showed signi ican bene icial changes in he lep in and
adiponec in concen a ions in he VLCHF g oup a e 12 weeks.
The signi ican dec ease in lep in le els also occu ed in he
VLCHF+HIIT g oup, whe eas he adiponec in inc ease was no
signi ican . No such bene icial changes we e de ec ed in he HIIT
and Con ol g oups. A low adiponec in-lep in a io has been
p oposed as a p omising ma ke o adipose issue dys unc ion
and may lead o ch onic sys emic in lamma ion. Lep in is
in ol ed in in lamma o y esponses, and i s inc eased le els
a e induced by adiposi y. In con as , a dec ease in adiposi y
leads o inc eased adiponec in le els, which is conside ed an
an i-in lamma o y ma ke (38,39).
We ha e al eady demons a ed a o able changes in se um
adiponec in and lep in concen a ions in heal hy young
indi iduals ollowing a compa able 12-week VLCHF die (16).
The p esen s udy shows ha a VLCHF die can induce simila
changes in indi iduals wi h o e a cons i u ion. Unlike ou
p e ious s udy wi h heal hy young indi iduals, in which body
mass educ ion was only small (16), he dec ease in lep in
and inc ease in adiponec in le els in his s udy migh be
F on ie s in Nu i ion | www. on ie sin.o g 9May 2022 | Volume 9 | A icle 867690