Content uploaded by Michelle Jospe
Author content
All content in this area was uploaded by Michelle Jospe on Jun 30, 2015
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
M E T H O D O L O G Y Open Access
Adherence to hunger training using blood
glucose monitoring: a feasibility study
M. R. Jospe
1
, R. C. Brown
1
, M. Roy
2
and R. W. Taylor
2*
Abstract
Background: “Hunger training”, which aims to teach people to eat only when blood glucose is below a set target,
appears promising as a weight loss strategy. As the ability of participants to adhere to the rigorous protocol has
been insufficiently described, we sought to determine the feasibility of hunger training, in terms of retention in the
study, adherence to measuring blood glucose, and eating only when blood glucose concentrations are below a set
level of 4.7 mmol/L.
Method: We undertook a two-week feasibility study, utilising an adaptive design approach where the specific blood
glucose cut-off was the adaptive feature. A blood glucose cut-off of 4.7 mmol/L (protocol A) was used for the first
20 participants. A priori we decided that if interim analysis revealed that this cut-off did not meet our feasibility
criteria, the remaining ten participants would use an individualised cut-off based on their fasting glucose
concentrations (protocol B).
Results: Retention of the participants in the study was 97 % (28/29 participants), achieving our criterion of
85 %. Participants measured their blood glucose before 94 % (95 % CI 91, 98) of eating occasions (criterion
80 %). However, participants following protocol A, which used a standard blood glucose cut-off of 4.7 mmol/L,
were only able to adhere to eating when blood glucose was below the prescribed level 66 % of the time, below our
within-person criterion of 75 %. By contrast, those participants following protocol B (individualised cut-off) adhered to
the eating protocol 84 % of the time, a significant (p= 0.010) improvement over protocol A.
Conclusion: Hunger training appears to be a feasible method, at least in the short-term, when an individualised fasting
blood glucose is used to indicate that a meal can begin.
Keywords: Food intake regulation, Hunger, Obesity, Blood glucose self-monitoring, Feasibility study, Adherence
Background
The persistent obesity epidemic has generated a plethora
of weight loss studies that investigate the effectiveness of
diets varying in macronutrient recommendations. How-
ever, it appears that modifying the composition of diets
has a minor impact on weight loss, especially over six
months or longer [1–5]. Instead, a far more relevant factor
appears to be the degree of adherence to the prescribed
diet [1, 6, 7], which may be encouraged by behavioural
strategies [8].
Oneeffectivestrategyforweightlossmaybelearningto
eat only when hungry, as eating in response to environmen-
tal, social, or emotional cues rather than physical hunger
has been consistently associated with a higher body mass
index (BMI) and energy intake [9–11]. Ciampolini et al.
[12] have developed an intriguing protocol that shows
promise in training people to eat according to their hunger.
Participants are trained to connect their physical symptoms
of hunger with their blood glucose, and to eat only when
their blood glucose is below a set target of 4.7 mmol/L.
When tested in a group of 74 overweight participants, this
method of “hunger training”(or “hunger recognition”,as
coined by Ciampolini et al. [12]) produced significantly
greater weight loss over 5 months (3.5 kg), compared with
that observed following the conventional approach of in-
creasing vegetable intake and physical activity [13].
While these initial results appear promising, replication
by another group would provide further support for the
use of this weight management strategy. The upcoming
* Correspondence: rachael.taylor@otago.ac.nz
2
Department of Medicine, University of Otago, PO Box 56, Dunedin 9054,
New Zealand
Full list of author information is available at the end of the article
© 2015 Jospe et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://
creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Jospe et al. Nutrition & Metabolism (2015) 12:22
DOI 10.1186/s12986-015-0017-2
SWIFT trial will test the effectiveness of four behavioural
strategies, including hunger training, on adherence to
diets and the resulting weight loss in overweight adults
during a two-year randomised controlled trial (RCT) [14].
However, we believed there were several important con-
siderations regarding the feasibility of the hunger training
protocol that required addressing before we could in-
corporate it into our larger trial. Firstly, adherence to the
protocol does not appear to have been reported, but is
critical given that even those who need to self-monitor
their blood glucose for diabetes management struggle to
do so even once per day [15]. Secondly, some concern was
raised regarding the suitability of 4.7 mmol/L as a stand-
ard cut-off for all participants given that this is below the
fasting glucose level for the majority of non-diabetic adults
[16, 17]. Lastly, there was some doubt over whether blood
glucose is an appropriate measure of perceived hunger,
given that different studies have indicated that blood glu-
cose and hunger are not correlated [18], whereas others
[19] have reported significant correlations as high as r =
0.55-0.63.
Given our concerns, we decided to undertake a feasi-
bility study [20] before including hunger training as a
support strategy in the larger SWIFT trial. We sought to
answer questions about adherence, the use of 4.7 mmol/L
as a cut-off, and the appropriateness of using blood glucose
to indicate hunger. We used an adaptive design method
[21], with the blood glucose cut-off as the adaptive feature
[22] to increase our chance of a finding a viable method
for hunger training. Adaptive design trials are formulated
to allow modification to the study while it is underway, in
order to efficiently learn from the data being gathered. A
priori rules guide the modification of the adaptive feature
(the characteristic of the protocol and study which
may require modification), and decisions are made based
on cumulative data from interim analyses [21, 22]. For this
study, we decided a priori to analyse the adherence data
to the original protocol (protocol A) after 20 participants,
before deciding on modifications to the blood glucose cut-
off based on predefined feasibility criteria.
The aim of our single-centre, single-arm study was to
determine the feasibility of the hunger training protocol
[12]–with particular attention to study retention, adher-
ence of measuring blood glucose, and adherence to eat-
ing below the blood glucose cut-off-to inform the design
and use of hunger training within the large randomised
controlled SWIFT trial [14].
Methods
Participants and setting
Participants were recruited by advertisement and word
of mouth in Dunedin, New Zealand. Eligible participants
of both sexes had to be at least 18 years of age. We re-
cruited both normal weight and overweight participants,
aiming for a final ratio of 2:1 (overweight: normal weight).
While only overweight participants will be included in
the main SWIFT trial, normal weight participants were
included in this feasibility study primarily to determine
whether the results vary with BMI and whether hunger
training might be suitable for weight maintenance in nor-
mal weight participants, as was suggested in the original
protocol [13]. Exclusion criteria included having type 1 or
type 2 diabetes or heart disease. Participants had to be
willing to measure their blood glucose via finger prick test
three to eight times per day, and to wear a continuous
glucose monitoring system if required. Participants
were informed about the experimental feasibility design
before written informed consent was obtained. The study
was approved by the University of Otago Human Ethics
Committee (H14/076).
Intervention
Participants followed the hunger training procedure for
two weeks, which was based on Ciampolini et al. [13].
Participants were provided with a booklet in which to rec-
ord hunger levels, blood glucose values, and food con-
sumed (Fig. 1). Every time a participant wanted to eat, they
were instructed to assess their hunger level using a visual
analogue scale ranging from 0 mm (not at all hungry) to
100 mm (extremely hungry) [23, 24]. Once that was com-
pleted, participants were instructed to measure capil-
lary blood glucose from a finger prick sample by portable
glucometer (Abbott Freestyle Optium Glucose Meter,
Australia). Participants were only permitted to eat if their
blood glucose was under their specified cut-off. If their
blood glucose was above their cut-off, they were instructed
to choose an activity that distracted them from food, and
to wait for new feelings of hunger for at least an hour be-
fore testing their blood glucose again (Fig. 2). Participants
were permitted to consume hypocaloric drinks at any time.
Participants were also specifically instructed that they
could only drink alcohol if their blood glucose was below
their cut-off.
We suggested that participants eat primarily fruit and
vegetables for the first few days of hunger training as per
Ciampolini et al. [13] in order to experience hunger more
frequently and therefore become more familiar with the
symptoms of hunger. As the training progressed, we en-
couraged participants to gradually reintroduce their typ-
ical foods and pay attention to the effect on hunger and
blood glucose.
Blood glucose Cut-off
To test the appropriateness and feasibility of the blood
glucose cut-off, we used an adaptive design approach
[21], with blood glucose cut-off as the adaptive feature
[22]. A priori we decided to initially use 4.7 mmol/L as
the cut-off (protocol A) as per the original protocol [12]
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 2 of 10
and to undertake an interim analysis after the first 20
participants had completed the training. If this analysis
indicated that the average within-person proportion of
eating occasions where blood glucose was below the cut-
off was more than 75 %, the cut-off would be individua-
lised following protocol B.
Protocol A
All participants in cohort A could only eat if their blood
glucose was 4.7 mmol/L or less.
Protocol B
Each participant in cohort B had an individual blood
glucose cut-off, which was calculated as the average of
the fasting glucose of the first two days of hunger train-
ing. If a participant’s individualised cut-off was less than
4.7 mmol/L, they were given a cut-off of 4.7 mmol/L.
During the first two days of hunger training, participants
used their fasting glucose of that morning for their cut-
off for the day.
Procedure
Participants met with the researcher on three occa-
sions over two weeks (baseline, day 7 and day 14). At
the first visit, height was measured with a fixed stadi-
ometer (Heightronic, QuickMedical, WA, USA) and weight
by electronic scales (Tanita BC-418). Participants com-
pleted a brief questionnaire that included demographic in-
formation [25], the self-administered short form of the
International Physical Activity Questionnaire (IPAQ) [26],
and the Intuitive Eating Scale-2 [27]. Participants were in-
troduced to hunger training and taught to measure their
blood glucose. After the first day of hunger training, the
researcher telephoned the participant to ensure that the
instructions were understood and able to be followed, and
to answer any questions. At the second visit participants
had the opportunity to ask questions and talk about any
challenges or successes. If participants had difficulty
only eating when their blood glucose was 4.7 mmol/L
or less in the first week, they wore a continuous glucose
monitoring system (CGMS) (iPro2 Continuous Glucose
Monitoring System, Medtronic, California, USA) for the
second week of the study. The CGMS allowed us to ob-
serve blood glucose variation in greater detail than the
intermittent finger prick glucose tests. The sensor was
inserted subcutaneously into the abdominal area and
monitored glucose in the interstitial fluid every five mi-
nutes for seven days. On the last visit, all participants
Fig. 1 The example page in the hunger training booklet
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 3 of 10
had their weight measured again, and were asked about
their experience of the study during a semi-structured
interview, which was recorded (Philips Voice Tracer
digital recorder LFH0622) and transcribed. The exit in-
terviews were examined with a data-driven thematic
analysis.
Feasibility criteria
A priori we specified that hunger training would be con-
sidered feasible if all of the following criteria were met:
1. 85 % or more of participants completed the study.
2. The average within-person proportion of eating
occasions where blood glucose was measured was
more than 80 %.
3. The average within-person proportion of eating
occasions where blood glucose was below the
cut-off was more than 75 %.
We were also interested in examining the within-person
correlation between hunger and blood glucose, and body
Fig. 2 Hunger training instructions in the hunger training booklet
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 4 of 10
weight change (if any) in overweight participants over the
two weeks.
Analyses
Descriptive statistics were used to answer the feasibility
criteria, with adherence measures and the correlation
between hunger and blood glucose calculated for each
participant before generating within-person point and
interval estimates. Study completion was calculated as
the number of participants who attended the final visit
at day 14 divided by the number of participants who
attended the first visit. Adherence to measuring blood
glucose was calculated by dividing the number of re-
ported eating occasions where blood glucose was noted by
the total number of eating occasions for each participant.
Adherence to eating when below the specified blood glu-
cose cut-off was calculated by dividing the number of
reported eating reported eating occasions where blood
glucose was below the assigned cut-off by the total number
of eating occasions with a noted blood glucose value. Dif-
ferences between groups were analysed using t-tests, with
unequal variances where indicated. Paired t-tests were
used to compare the average of the first two days of fast-
ing glucose with fasting glucose for the 14-day period.
As no formal power calculations are undertaken in feasi-
bility studies, sample sizes should be based on estimating
feasibility outcomes (Arain, Campbell, Cooper, & Lancaster,
2010). We estimated that a minimum of 25 participants
was appropriate to estimate the retention rate and adher-
ence to the hunger training intervention. Statistical ana-
lysis was performed using Stata 12 (StataCorp, College
Station, TX).
Results
Recruitment took place on July 2, 2014 by sending an in-
vitation email to staff and students at the University of
Otago in Dunedin, New Zealand. All participants were
recruited in a single day. The pilot study took place from
July 8 to Aug 6, 2014 and 30 participants were recruited.
One participant discontinued hunger training due to
fasting glucose levels being above the cut-off for possible
diabetes diagnosis (>7 mmol/L), resulting in 29 partici-
pants being included in the final analysis.
The participants were predominantly well-educated,
white women, with an average BMI of 31 kg/m
2
(Table 1).
As there were no significant differences in demographic
characteristics at baseline between cohorts A and B (data
not shown), only the combined results are presented.
Adherence
Adherence was measured in terms of overall study reten-
tion, glucose measurement prior to eating, and compli-
ance to the blood glucose cut-off. There was no difference
in any adherence measures between lean and overweight
participants (data not shown). Retention was high, with 28
out of 29 participants (97 %) completing the pilot study,
well above our predetermined criterion for determining
success of 85 %.
In terms of adherence to the blood testing protocol,
the within-person proportion for measuring blood glu-
cose before eating was 94 % (95 % CI 91, 98) of eating
occasions, with no significant difference between cohorts
A and B. This level of adherence considerably exceeded
our a priori requirement of measuring before 80 % of
eating occasions.
Adherence to eating with protocol A
Participants following protocol A, which used a univer-
sal blood glucose cut-off of 4.7 mmol/L, adhered to the
goal of eating only when blood glucose was below this
level 66 % of the time, which was below our within-
person adherence requirement of 75 % (Table 2). Fur-
thermore, four out of the nineteen (21 %) participants in
this cohort adhered to protocol A on less than half of
their eating occasions (Fig. 3), with one participant only
adhering on two eating occasions over the two week
period (5 % of all eating occasions).
Of these 19 participants, six who appeared to be strug-
gling with only eating when their blood glucose was
4.7 mmol/L or less during the first week of training wore
a continuous glucose measurement sensor during week
Table 1 Characteristics of participants
Variable All (N= 29)
Age (years) 43.3 ± 12.5
Height (m) 1.7 ± 0.1
Weight (kg) 89.8 ± 25.2
Body mass index (kg/m
2
) 31.2 ± 9.0
Women, n (%) 22 (76 %)
White ethnicity, n (%) 27 (93 %)
Partnership status, n (%)
Partnered 16 (55 %)
Non-partnered 13 (45 %)
University degree, n (%) 19 (66 %)
Intuitive eating questionnaire
Overall score
a
3.1 ± 0.5
Unconditional permission to eat 3.3 ± 0.6
Eating for physical rather than emotional reasons 2.9 ± 0.7
Reliance on hunger 2.9 ± 0.7
Body-food choice congruence 3.4 ± 0.9
Physical activity (MET-minutes/week)
b
2019 ± 2955
Hours of sitting per day
b
7.1 ± 2.6
Note: Un less indicated, values are mean ± SD
a
The intuitive eating score ranges from one to five, with higher scores
indicating greater levels of intuitive eating
b
median ± IQR
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 5 of 10
two. On average, their blood glucose was above the proto-
col A cut-off of 4.7 mmol/L 85 % of the time over the
seven days, or 143 hours of the total 168 hours captured.
Additionally, five of these six participants had at least one
day where their blood glucose never dropped below
4.7 mmol/L. Fig. 4 illustrates the results from one of these
participants, showing their blood glucose over 7 days and
the amount of time it was under both protocol A (univer-
sal 4.7 mmol/L cutoff) and protocol B (individualised cut-
off of 6.2 mmol/L for this participant).
Adherence to eating with protocol B
A priori we decided that the blood glucose cut-off would
be deemed feasible if the within-person proportion of
measured blood glucose was below the cut-off on more
than 75 % of eating occasions. As protocol A was below this
requirement, we switched to protocol B for the remaining
ten participants. For protocol B participants, their indivi-
dualised blood glucose cut-off was calculated from the aver-
age of fasting glucose over the first two days of hunger
training, which resulted in an average cut-off of 5.5 mmol/L
(SD 3.0). The use of this individualised protocol increased
adherence to the eating protocol to a within-person pro-
portion of 84 % of the time, which was a significant
increase over protocol A (p= 0.010) and met our bench-
mark for feasibility.
We examined whether the average of the first two days
of fasting glucose was a suitable estimate of the average
fasting glucose throughout the two-week period, and there-
fore was suitable to determine the individualised cut-off.
Fasting glucose means were comparable (difference
0.07 mmol/L), with no significant difference between
these time periods (p= 0.415).
Correlation between hunger and blood glucose
A significant, albeit modest, inverse within-person cor-
relation was observed between perceived hunger and blood
glucose concentrations of r = −0.23 (95 % CI −0.15, −0.31),
p< 0.001. Cohort B tended to have a stronger correlation
between hunger and glucose (r = −0.33 for cohort B vs
r=−0.18 for cohort A) although our small numbers
probably preclude statistical significance (p= 0.082).
Weight loss
Overweight participants following both protocols achieved
significant weight loss over the two-week period, with an
average loss of 1.5 kg (95 % CI 2.2, 0.9) and a correspond-
ing reduction in BMI of 0.6 kg/m
2
(95 % CI 0.3, 0.8), p<
0.001 (Table 3). By contrast, lean participants maintained
Table 2 Adherence to hunger training
Adherence Cohort A (N= 19) Cohort B (N= 10) p-value
Measuring glucose 92.3 % (88.6, 96.0) 96.1 % (93.9, 98.2) 0.149
Eating below glucose cut-off 66.4 % (54.2, 78.6) 84.4 % (78.0, 90.8) 0.010
Data presented as within-person mean (95 % CI)
Adherence to Cutoff (%)
Percent of Total Participants in Cohort
0
20
10
40
30
60
70
50
0 20406080100
Cohort A
0 20406080100
Cohort B
Fig. 3 Histogram of adherence to eating below the blood glucose cut-off in cohort A and cohort B
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 6 of 10
their weight (p= 0.337, data not shown). Although partici-
pants following protocol B appeared to lose more weight
than those in protocol A, differences were not significant
(Table 3).
Exit interviews
The major themes that emerged from the exit interviews
were: awareness of non-hungry eating, change of dietary
intake, and the use of blood glucose monitoring.
Awareness of non-hungry eating
Participants reflected that hunger training made them
realise that they were used to eating for reasons other
than hunger, including as a way to stave off boredom, to
cope with emotions, and because of habit:
It certainly does make you more aware–it was a
good thing to do. It cut down a lot of my night-time
snacking, just cruising past and something goes in
my mouth without thinking. That’smymajor
problem, this night-time grazing.
It’s well up there because it’s something that I don’t
really pay attention to. I’m just an automatic feeder.
Needing to stop and think before I put something in
my mouth was really good. And I very seldom eat
when I’m hungry. I’m an emotional eater: bored,
tired, etc.…I really want to stop doing that.
Change of Dietary Intake
Most participants reduced or eliminated snacks:
I stopped eating morning tea–just a habit because
everyone else has something to eat. But I mean, I
have an office job–I’m just sitting on my bottom
not burning up much energy, so now I just have
coffee instead. By lunchtime now, I’m feeling
hungry.
The thing I cut out was the snacking–that was the
main impact. For instance I used to eat bag of chips
for no reason, just because it seemed like a good
idea.
Many participants chose better food, because they were
more conscious of their food intake or because they no-
ticed how their blood glucose reacted afterwards:
My lunches I’ve changed totally. It’s now three
mandarins, a pottle of yoghurt, and a banana. And
before? It was anything that I wanted.
I’m not eating things like chips and crackers. Now
it’s a treat. It’s really changed how much processed
food and I’m definitely eating much more
vegetables, and enjoying them. It just seems like a
waste now to eat a bag of chips, since then I can’t
have dinner.
Many participants reduced their portion sizes at meals,
and noted that they were surprised that they didn’t need
as much food as they had previously assumed, and could
no longer imagine eating as much as before the pilot:
I still served myself the same portions but I couldn’t
eat it all. Which is kind of weird for me. I stopped
when I had enough. I don’t know if I’m just getting
used to eating a bit less or I realise that I didn’t
NEED to eat it all. Whereas before I would finish it
because it was wasteful. But now I think, “you don’t
have to force yourself to eat stuff that you don’t
need to.”And that rating of fullness at the end was
quite good, because before I would have that too
full, gross, feeling. But now thinking about what full
feels like and overfull feels like and what not quite
full feels like.
Cutting down on what I was having for breakfast
did make a difference to my blood sugar before
lunchtime. So just having porridge OR toast rather
than porridge AND toast meant that I could have
lunch at lunchtime.
Day 1
(partial day) Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Average
Fig. 4 Results from the continuous glucose monitoring sensor from a participant over 7 days, comparing the amount and percentage of time the
participant was below the protocol A cut-off of 4.7 mmol/L and protocol B individualised cut-off of 6.2 mmol/L
Table 3 Change in weight and BMI over 2 weeks in overweight
participants
Variable Cohort A (N= 13) Cohort B (N=9) p-value
Weight change (kg) −1.1 (−0.5, −1.8) −2.1 (−0.6, −3.6) 0.134
BMI change (kg/m
2
)−0.4 (−0.2, −0.6) −0.8 (−0.2, −1.4) 0.137
Data presented as mean (95 % CI)
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 7 of 10
Conversely, a few participants reported that their por-
tion sizes increased, since they were hungrier than usual
because they were eating less frequently:
But when I was hungry, I ate MORE. Because I was
having to wait until lunchtimes until I could eat, so
when I got lunch, I ate more than I normally would
have. If I hadn’t eaten until 1 pm, anything that
wasn’t tied down wasn’t safe! Once I pricked my
finger, I would have my meal and then the
chocolate, because I better have the chocolate now
else I’ll have to stab my finger again. It eliminates
the grazing but increased my portion sizes.
The use of blood glucose monitoring
Many participants commented that their blood glucose
was an unpredictable measure of their hunger:
I was having trouble with the [glucose] readings and
matching it up with what I was doing, but it
certainly made me think about registering whether I
was hungry rather than just eating because of
routine.
Super hungry didn’t seem to corresponding to
particularly low glucose. I did find it frustratingly
inaccurate in terms of measuring my hunger,
even though I was much more in touch with
my hunger.
However, most participants viewed measuring their blood
glucose as a useful behaviour for gaining awareness of their
eating habits:
It’s not just the fact that you inflict pain on yourself
–it’s the fact that you inflict pain and it might say
“no”anyway. Really have to think, “look, am I
actually feeling hungry enough?”. I think it’s
extremely effective since it just makes you more
aware. Even if I didn’t hurt, it makes you aware of
those wee niggle, but “I’ve only eating a little awhile
ago”. The pain element is useful and the fear of the
rejection after the pain.
The psychological thing of having to prick your
finger every time you want to eat is a bit of a red
herring but it’s quite a relevant thing. I think “hmm,
I would like some afternoon tea but my fingers are a
bit sore today…maybe I don’t need it”
Many participants commented that the pain with finger
pricking reduced after the first week:
I have the feeling that stabbing gets better. At the
beginning it hurt more, but I really can’t feel it
anymore.
I don’t find it particularly onerous. It’s easy enough
to fit and only a minor irritation, to prick your
finger.
Discussion
The idea that using finger prick blood glucose monitoring
to train individuals to eat only when hungry appeared to
be a promising method for weight management, based on
the results from a group of researchers [12]. However, nei-
ther this original study, nor an offshoot from these authors
[28] reported on adherence, a critical consideration for
determining success. Thus we repeated the Ciampolini
protocol (protocol A) using an adaptive design, which
allowed us to determine if an alternative blood glucose cut-
off might be more suitable. In fact, we found that only our
adaptive version of hunger training (protocol B) was feas-
ible, in terms of meeting all three criteria: study retention,
adherence to measuring blood glucose, and adherence to
eating below the blood glucose cut-off. It is difficult to
compare our findings as neither of the previous studies
reported on feasibility per se. However, both previous
studies reported retention of 80 % of participants [13, 28],
which is likely comparable to our retention rates, consid-
ering their longer duration.
We anticipated that adherence to measuring blood
glucose prior to every eating occasion would be an issue,
however this did not seem to be the case. Participants
measured their blood glucose before nearly every meal
over this two-week trial. Our qualitative data suggests
that participants thought that the pain and hassle of
pricking their fingertips and measuring their blood glu-
cose was not as bad as they had originally expected. Al-
though our adherence to measuring blood glucose was
considerably higher than that reported with diabetics
[15], this is likely to be influenced by the short-term na-
ture of our trial.
While both the original (protocol A) and adapted
(protocol B) protocols achieved our criteria for success
of feasibility in terms of retention and adherence to
measuring blood glucose, only the protocol B met the
criterion for adherence to eating below the blood glu-
cose cut-off. Participants’difficulty of adhering to only
eating when their blood glucose was under 4.7 mmol/L
is consistent with large surveys that show that this con-
centration is below the fasting glucose level for the ma-
jority of non-diabetic adults [16, 17]. Thus adapting our
protocol to use an individualised blood glucose cut-off
substantially improved adherence to only eating when
blood glucose was below the cut-off. Furthermore, this
adaptation might improve weight loss success as well, at
least over the short-term. Our overweight participants
lost 1.5 kg on average over the fortnight, with those fol-
lowing protocol B losing nearly twice as much weight
compared with protocol (although this difference did
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 8 of 10
not reach statistical significance). This finding was intri-
guing given we anticipated that the more lenient cut-off
(protocol B) would diminish rather than enhance weight
loss. However, both the diary information and the ac-
companying qualitative interviews showed that protocol
B was more successful at encouraging participants to
only eat when hungry. One limitation of our data is that
although we requested brief dietary information as part
of the two-week diary, this was not detailed enough to
allow energy intake to be calculated. Thus exactly how
this more lenient cut-off changed dietary intake is uncer-
tain. Furthermore our trial only lasted two weeks and
longer follow-up is required to determine whetherhun-
ger training offers a viable method for sustained weight
loss. As hunger training will be one support strategy in
the upcoming SWIFT trial, we will have the opportunity
to examine the effectiveness of hunger training on weight
loss over two years. However, our findings do support
those of Ciampolini et al. who demonstrated an average
weight loss of 5.8 kg over 5 months in overweight partici-
pants [13].
Our findings of only a weak correlation between hunger
and blood glucose are in agreement with the range of re-
sults reported in the literature [18, 19]. The weak and in-
consistent association between hunger and blood glucose
was the most problematic element for participants, as evi-
dent in their feedback. However, most participants felt
that measuring blood glucose still provided valuable feed-
back and was crucial for modifying their eating behaviour,
irrespective of its limitations. It seems as if the benefits of
objective biofeedback provided by measuring their blood
glucose levels enhances the benefits of monitoring appetite
[29, 30] and dietary intake [31]. Similarly, self-monitoring
of blood glucose in diabetics has been shown to improve
adherence to nutritional recommendations and decrease
body weight [32]. Furthermore, the necessity to measure
blood glucose concentrations required participants to rec-
ord their appetite and blood glucose results at that in-
stant, rather than hours later, which may be beneficial
as a shorter time interval between eating and dietary self-
monitoring has been shown to be significantly associated
with weight loss [33].
The main strength of this feasibility study is our care-
ful measure of adherence, which has not previously been
described for this intriguing weight management strat-
egy. Feasibility and pilot studies are an important step
before a large trial, and are especially useful when feasi-
bility objectives and success criteria are defined a priori
[34]. Including an adaptive design approach helped us to
efficiently find the most suitable blood glucose cut-off
for our participants, which appears to be an improve-
ment over the original method. This study provided
us with confidence about the feasibility of the hunger
training method in general, and knowledge that an
individualised blood glucose cut-off is the most viable
approach. Hunger training may provide an effective
strategy for weight loss by teaching people to eat accord-
ing to their physical hunger rather than in response to
environmental, social, or emotional cues. Based on our
findings, the SWIFT trial will include our adapted ver-
sion of hunger training as one of four intervention arms
[14]. However, our study also has some limitations. As
befitting a feasibility study, it was relatively short, but
was based on the original protocol that showed the ma-
jority of participants were trained after two weeks. While
it would have been interesting to use continuous monitor-
ing in all participants, it was cost prohibitive. However,
with the advancements in non-invasive glucose monitor-
ing, such as using contact lenses [35] and temporary tat-
toos [36], continuous (and pain-free) blood glucose
monitoring is likely to become more affordable and
accessible.
Conclusions
Our results show that participants are willing to partici-
pate in hunger training, measure their glucose before eat-
ing, and only eat when their blood glucose is under their
individualised cut-off, at least in the short-term. The re-
sults of our study suggest that the adapted version of the
hunger training protocol is feasible, and viable to use as a
support strategy in the SWIFT trial. Further testing of
hunger training in the SWIFT trial will allow us to exam-
ine the effects of this method over a two-year duration in
combination with healthy eating and exercise advice.
Abbreviations
BMI: Body mass index; CGMS: Continuous glucose monitoring system;
CI: Confidence interval; IQR: Interquartile range; MET: Metabolic equivalent of
task; SD: Standard deviation.
Competing interests
The authors declare that they have no competing interests.
Authors’contributions
MJ carried out the feasibility study, including all participant interactions,
analysed the data, and drafted the manuscript. RB participated in the design
and coordination of the study and helped to draft the manuscript. MR
participated in the design of the study, provided medical oversight, and
helped to draft the manuscript. RT conceived of the study, and participated
in the design and coordination and helped to draft the manuscript. All of
the authors have been involved in the writing of this manuscript and have
read and approved the final text.
Acknowledgements
We wish to thank the University of Otago for funding, Dr Jill Haszard and
Dr Josie Athens for statistical advice, and all the participants who
volunteered their time and fingertips to this study.
Author details
1
Department of Nutrition, University of Otago, Dunedin, New Zealand.
2
Department of Medicine, University of Otago, PO Box 56, Dunedin 9054,
New Zealand.
Received: 17 March 2015 Accepted: 29 May 2015
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 9 of 10
References
1. Johnston BC, Kanters S, Bandayrel K, Wu P, Naji F, Siemieniuk RA, et al.
Comparison of weight loss among named diet programs in overweight and
obese adults: a meta-analysis. JAMA. 2014;312:923–33.
2. Wycherley TP, Moran LJ, Clifton PM, Noakes M, Brinkworth GD. Effects of
energy-restricted high-protein, low-fat compared with standard-protein,
low-fat diets: a meta-analysis of randomized controlled trials. Am J Clin Nutr.
2012;96:1281–98.
3. Ajala O, English P, Pinkney J. Systematic review and meta-analysis of different
dietary approaches to the management of type 2 diabetes. Am J Clin Nutr.
2013;97:505–16.
4. Atallah R, Filion KB, Wakil SM, Genest J, Joseph L, Poirier P, et al. Long-term
effects of 4 popular diets on weight loss and cardiovascular risk factors: a
systematic review of randomized controlled trials. Circ Cardiovasc Qual
Outcomes. 2014;7:815–27.
5. Dutton GR, Laitner MH, Perri MG. Lifestyle interventions for cardiovascular
disease risk reduction: a systematic review of the effects of diet composition,
food provision, and treatment modality on weight loss. Curr Atheroscler Rep.
2014;16:442–14.
6. Dansinger ML, Gleason JA, Griffith JL, Selker HP, Schaefer EJ. Comparison of
the Atkins, Ornish, Weight watchers, and Zone diets for weight loss and
heart disease risk reduction. JAMA. 2005;293:43–53.
7. Pagoto SL, Appelhans BM. A call for an end to the diet debates. JAMA.
2013;310:687–8.
8. Michie S, Abraham C, Whittington C, McAteer J, Gupta S. Effective
techniques in healthy eating and physical activity interventions:
a meta-regression. Health Psychol. 2009;28:690–701.
9. French SA, Epstein LH, Jeffery RW, Blundell JE, Wardle J. Eating behavior
dimensions. Associations with energy intake and body weight. A review.
Appetite. 2012;59:541–9.
10. French SA, Mitchell NR, Wolfson J, Finlayson G, Blundell JE, Jeffery RW.
Questionnaire and laboratory measures of eating behavior. Associations
with energy intake and BMI in a community sample of working adults.
Appetite. 2014;72:50–8.
11. Baños RM, Cebolla A, Moragrega I, Van Strien T, Fernández-Aranda F,
Agüera Z, et al. Relationship between eating styles and temperament in an
Anorexia Nervosa, Healthy Control, and Morbid Obesity female sample.
Appetite. 2014;76:76–83.
12. Ciampolini M, Bianchi R. Training to estimate blood glucose and to form
associations with initial hunger. Nutr Metab. 2006;3:42.
13. Ciampolini M, Lovell-Smith D, Sifone M. Sustained self-regulation of energy
intake. Loss of weight in overweight subjects. Maintenance of weight in
normal-weight subjects. Nutr Metab. 2010;7:4.
14. Taylor RW, Roy M, Jospe MR, Osborne HR, Meredith-Jones K, Williams SM
et al. Determining how best to support overweight adults to adhere to
effective lifestyle change: Protocol for the SWIFT study. under review.
15. Coyle ME, Francis K, Chapman Y. Self-management activities in diabetes
care: a systematic review. Austral Health Rev. 2013;37(4):513–22.
16. McAuley KA, Williams SM, Mann JI, Goulding A, Chisholm A, Wilson N, et al.
Intensive lifestyle changes are necessary to improve insulin sensitivity:
a randomized controlled trial. Diabetes Care. 2002;25:445–52.
17. Rapp K, Schroeder J, Klenk J, Ulmer H, Concin H, Diem G, et al. Fasting
blood glucose and cancer risk in a cohort of more than 140,000 adults in
Austria. Diabetologia. 2006;49:945–52.
18. Flint A, Gregersen NT, Gluud LL, Møller BK, Raben A, Tetens I, et al.
Associations between postprandial insulin and blood glucose responses,
appetite sensations and energy intake in normal weight and overweight
individuals: a meta-analysis of test meal studies. Br J Nutr. 2007;98:17–25.
19. Lemmens SG, Martens EA, Kester AD, Westerterp-Plantenga MS. Changes in
gut hormone and glucose concentrations in relation to hunger and fullness.
Am J Clin Nutr. 2011;94:717–25.
20. National Institute for Health Research: Glossary. www.nihr.ac.uk/glossary/.
Accessed 10 Feb 2015.
21. Chow SC, Chang M. Adaptive design methods in clinical trials-a review.
Orphanet J Rare Dis. 2008;3:11–24.
22. Lorch U, O’Kane M, Taubel J. Three steps to writing adaptive study
protocols in the early phase clinical development of new medicines. BMC
Med Res Methodol. 2014;14:84.
23. Blundell J, De Graaf C, Hulshof T, Jebb S, Livingstone B, Lluch A, et al.
Appetite control: methodological aspects of the evaluation of foods. Obes
Rev. 2010;11:251–70.
24. Stubbs RJ, Hughes DA, Johnstone AM, Rowley E, Reid C, Elia M, et al. The
use of visual analogue scales to assess motivation to eat in human subjects:
a review of their reliability and validity with an evaluation of new hand-held
computerized systems for temporal tracking of appetite ratings. Br J Nutr.
2000;84:405–15.
25. Statistics New Zealand. www.stats.govt.nz. Accessed Feb 8 2015.
26. Craig CL, Marshall AL, Sjöström M, Bauman AE, Booth ML, Ainsworth BE,
et al. International physical activity questionnaire: 12-country reliability and
validity. Med Sci Sports Exerc. 2003;35:1381–95.
27. Tylka TL, Kroon Van Diest AM. The Intuitive Eating Scale-2: item refinement
and psychometric evaluation with college women and men. J Couns Psychol.
2013;60:137–53.
28. Ledoux T, Gallagher MR, Ciampolini M, Sampson M. Biofeedback enhanced
lifestyle intervention: exploring the experience of participants in a novel
intervention for disinhibited eating and obesity. OJPM. 2014;04:779–88.
29. Allen HN, Craighead LW. Appetite monitoring in the treatment of Binge
Eating Disorder. Behav Ther. 1999;30:253–72.
30. Dicker SL, Craighead LW. Appetite-focused cognitive-behavioral therapy in
the treatment of binge eating with purging. Cogn Behav Pract.
2004;11:213–21.
31. Burke LE, Wang J, Sevick MA. Self-monitoring in weight loss: a systematic
review of the literature. J Am Diet Assoc. 2011;111:92–102.
32. Duran A, Martin P, Runkle I, Perez N, Abad R, Fernandez M, et al. Benefits of
self-monitoring blood glucose in the management of new-onset Type 2
diabetes mellitus: the St Carlos Study, a prospective randomized clinic-based
interventional study with parallel groups. J Diabetes. 2010;2:203–11.
33. Burke LE, Sereika SM, Music E, Warziski M, Styn MA, Stone A. Using
instrumented paper diaries to document self-monitoring patterns in weight
loss. Contemp Clin Trials. 2008;29:182–93.
34. Thabane L, Ma J, Chu R, Cheng J, Ismaila A, Rios LP, et al. A tutorial on pilot
studies: the what, why and how. BMC Med Res Methodol. 2010;10:1.
35. Farandos NM, Yetisen AK, Monteiro MJ, Lowe CR, Yun SH. Contact Lens
Sensors in Ocular Diagnostics. Advanced Healthcare Materials 2014:n/a–n/a.
36. Bandodkar AJ, Jia W, YardımcıC, Wang X, Ramirez J, Wang J. Tattoo-based
noninvasive glucose monitoring: a proof-of-concept study. Anal Chem.
2014;87:394–8.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit
Jospe et al. Nutrition & Metabolism (2015) 12:22 Page 10 of 10
