Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.

668-94

Wien et al. Nutrition Journal 2013, 12:155

http://www.nutritionj.com/content/12/1/155

RESEARCH Open Access

A randomized 3x3 crossover study to evaluate the

effect of Hass avocado intake on post-ingestive
satiety, glucose and insulin levels, and subsequent
energy intake in overweight adults
Michelle Wien1*, Ella Haddad1, Keiji Oda2 and Joan Sabaté1,2

Abstract
Background: The behavioral outcome of food ingestion is a complex process that involves psychological and
biological factors. Avocados are nutrient dense with properties that may favorably impact energy balance. This
study sought to evaluate if incorporating approximately one half of a Hass avocado by addition or inclusion into a
lunch meal will influence post-ingestive satiety, glucose and insulin response, and subsequent energy intake among
overweight adults.
Methods: This was a randomized 3x3 single-blind crossover design study with 26 healthy overweight adults
(mean ±SD age 40.8±11.0 years and BMI 28.1±2.4 kg/m2). Participants consumed a standardized breakfast
followed by 1 of 3 lunch test meals [Control (C), avocado-free; Avocado Inclusive (AI); and, Avocado Added (AA)].
Participants rated five appetite sensations using a visual analog scale (VAS) before lunch and at specific intervals
over 5 hours following the start of the test meal. Blood glucose and insulin were measured before lunch and
at specific intervals over 3 hours following the start of the test meal. Mixed models were used to compare
differences among the 3 test meals, and the area under the curve (AUC0-xh) was computed for the VAS and
biological measures.
Results: There were significant differences in the AUC(0-5h) for the self-reported feelings of satisfaction (P=0.04)
and desire to eat (P=0.05) in the mixed model analysis. Compared to the C test meal, the AA test meal increased
satisfaction by 23% (P=0.05) and decreased the desire to eat by 28% (P=0.04) for the AUC(0-5h). For the AUC(0-3h), the AA
test meal increased satisfaction by 26% (P=0.02) and decreased the desire to eat by 40% (P=0.01) as compared
to the C test meal. Compared to the AI meal, the AUC(0-3h) for blood insulin was higher in the C and AA meals
(P=0.04 and P=0.05, respectively).
Conclusions: The addition of approximately one half of a Hass avocado at a lunch meal can influence post-ingestive
satiety over a subsequent 3 and 5 hour period in overweight adults. A caveat to these findings is that the avocado
contained an additional 112 kcal, which may have accounted for the observed increase in satisfaction and decreased
desire to eat. Future trials are warranted to evaluate the effects of avocado intake on weight management in adults of
varying BMIs and among insulin resistant individuals.
Keywords: Avocado, Satiety, Overweight, Glucose, Insulin, Visual analog scale, Dietary compensation

* Correspondence: [email protected]
1
Department of Nutrition, School of Public Health, Loma Linda University,
24951 N. Circle Dr., Nichol Hall 1102, Loma Linda, CA 92350, USA
Full list of author information is available at the end of the article

© 2013 Wien et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 2 of 9

http://www.nutritionj.com/content/12/1/155

Background Methods
The behavioral outcome of food ingestion is a complex We conducted a randomized 3x3 single-blind crossover
process that involves psychological and biological factors design study (three 1-day study periods scheduled 1 week
that culminates in an individual’s overall 24-hour energy in- apart) at Loma Linda University, Loma Linda, California.
take [1]. One of the components of the appetite system is Using a within subject repeated measures design, we
satiety, which reflects a process that leads to increased full- evaluated the effect of avocado intake on the short-term
ness after a meal, a decline in hunger, and inhibition of fur- regulation of food by employing the use of one of three
ther eating in the postprandial period. lunch test meals within a single day on three different
In addition to sensory quality, the macronutrient days. Each participant received all treatments on the
composition, physical structure and energy density of a same day of the week and had a 1 week washout period
whole food may contribute to the modulation of sati- between treatments.
ety. More recent research has shown that the volume
of a meal also influences satiety [2,3]. Additionally, the Eligibility criteria
quality of the fat composition, i.e. degree of saturation Healthy overweight and moderately obese adults were
of fatty acids in a food, may influence rates of oxidation recruited through the use of posters, flyers, and news-
and thermogenesis in animals and humans [4,5]. Fur- paper advertisements on the Loma Linda University
thermore, a single food may favorably impact energy campus and in the surrounding communities. A study
balance according to its ability to offset spontaneous web page was developed with a complete description of
energy reduction at the next meal(s), which is known the study and online application form. Applications were
as the dietary compensation score [6]. also taken by phone.
The fresh pulp weight of Hass avocados is 72% water Eligibility criteria were: age 25–65 years, body mass
and contains only 1.7 kcal/g, therefore they are clas- index (BMI, kg/m2) ≥25 and ≤35, weight stable for at
sified as a medium energy dense food (defined as a least 6 months, normoglycemic, normotensive, sedentary
range between 1.5-4.0 kcal/g) [7]. Hence, when added or low level of habitual activity (less than 10 hours of ex-
to a meal they will increase the volume similar to ercise per week), non-smoker, not dependent on caffeine,
other fruits and vegetables, which are food categories and not taking any medications known to influence
that have been previously shown to have a beneficial postprandial glucose and insulin levels. The recruitment
effect on weight control [8]. Further, the typical serv- process yielded 80 applicants and 56 individuals met the
ing size is approximately one half of a medium size av- eligibility criteria. Forty-seven individuals attended infor-
ocado (70 g) [9], which translates to being an excellent mation meetings and 30 were selected, plus 2 alternates.
source of dietary fiber (5 g). Fiber is another food Four of the selected applicants declined participation
component strongly linked to enhancement of satiety due to unforeseen scheduling conflicts. Both alternates
[10] and modulation of the glucose and insulin re- were included to achieve the targeted accrual of 28 par-
sponses to meals [11]. ticipants and goal of 25 completers for adequate power
The connection between potential biological markers (see Statistical Methods). One participant withdrew on
of appetite regulation continues to be an active area of the first day and one participant was asked to leave the
research in normal weight and overweight populations. second week due to non-compliance with the study
A 2007 meta-analysis by Flint et al. [12] has shown that protocol. The study was approved by the Loma Linda
the biological signaling of satiety by insulin in the over- University Institutional Review Board and informed writ-
weight population is blunted, which could lead to the ten consent was obtained from all participants.
deleterious consequence of overeating at subsequent ad
libitum meals and snacks. In light of the increased Anthropometric measurements
prevalence in overweight in humans and cross-sectional Height was measured to the nearest centimeter using a
evidence showing an increase in snacking and total en- stadiometer on the first study day. Weight was measured
ergy intake in the United States over the past three de- using an internally calibrated segmental body compos-
cades [13], the addition of approximately one half of an ition scale/analyzer (model TBF-300A, Tanita®, Arlington
avocado at a specific meal(s) may be a simple dietary Heights, IL) and recorded to the nearest 0.01 pound.
intervention to consider for individuals that consume BMI was calculated as weight(kg)/height(m2). The daily
excessive energy during specific snack and/or meal energy needs for each participant were estimated using
times. The aims of the present study are to evaluate if the Harris-Benedict equation after adjustment for over-
incorporating ~1/2 of a fresh Hass avocado by addition weight status, which was subsequently multiplied by an
or inclusion into a lunch meal will influence post- activity factor of 1.3 for sedentary lifestyle. Participants
ingestive satiety, the glycemic and insulin response, and were then assigned to receive either a 1600, 2000 or
subsequent energy intake in overweight adults. 2400 kcal daily meal plan.
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 3 of 9

http://www.nutritionj.com/content/12/1/155

Study meals and dietary compensation. The number of portions of

Under the direct supervision of trained study personnel in food items selected and consumed by the participants was
the Loma Linda University Department of Nutrition directly observed and written in a discreet manner by a se-
Metabolic Kitchen, participants consumed the same stan- nior investigator and trained research staff. To supplement
dardized breakfast meal containing 25% of their estimated the written documentation of the food items taken and
daily energy needs on each of the 3 assigned study days. leftovers remaining on the plate, a hidden video camera
For lunch, participants consumed 1 of 3 test meals: Con- was utilized to record the foodservice delivery process.
trol (C), avocado-free; Avocado Inclusive (AI); or, Avocado The leftovers were photographed using a digital camera
Added (AA) (see below). The standardized breakfast and and weighed to the nearest gram using a digital scale. Two
the lunch test meals were designed to deliver the recom- research assistants separately compared the written docu-
mended levels of macronutrients according to the Accept- mentation with the videotape recording, still photos and
able Macronutrient Distribution Ranges developed by the weight of leftovers to produce a record of food intake for
Food and Nutrition Board of the Institute of Medicine each participant. If any discrepancy existed between the
[14]. The participants consumed 13-14% energy from pro- two researcher’s records, a senior investigator reviewed all
tein, 49-51% energy from carbohydrate, and 35-38% en- of the data sources to determine the most valid measure-
ergy from fat at the three lunch test meals. Further, the C ment of dietary intake.
and AI lunch meals delivered 35% of the participant’s daily Pre-portioned evening snacks were provided to partici-
energy needs and the AA lunch meal provided 41% of the pants at the conclusion of the ad libitum dinner meal and
daily energy needs (Table 1). All foods were precisely mea- participants were asked to record any snacks consumed
sured or weighed to the nearest gram using a digital scale after leaving the research kitchen until going to bed or until
and the meals were matched for taste and appearance. midnight. Participants were contacted by phone the follow-
Participants were permitted to drink water with and be- ing morning by study personnel for a self-report of the in-
tween meals on the three study days according to their take of pre-portioned evening snacks. The energy and
typical pattern of water intake. macronutrient intake subsequent to the lunch test meals
The standardized breakfast meal included orange juice, was assessed based on the observed food consumed at the
cornflakes, milk and a commercially prepared scone. The ad libitum dinner meal and from the participant’s self-
C test meal included a salad (mixed greens, cherry toma- reported consumption of pre-portioned evening snacks.
toes, reduced fat Swiss cheese, Italian salad dressing), a re-
fined grain French baguette and commercial chocolate Visual analog scales
chip cookies. Fresh, ripe Hass avocados (provided by the By means of a mark on 100 mm line visual analog scales
Hass Avocado Board) were sliced and included or added to (VAS), participants rated their appetite sensations (hun-
the C test meal to produce the AI and AA test meals, re- ger, fullness, satisfied, desire for a meal, and prospective
spectively. The amount of avocado varied (range of ~50 to food consumption). The VAS was completed before
90 g) with the energy needs of the participant [75 g (~1/2 lunch and at approximately 30, 60, 90, 120, 180 and
of an avocado) for the 2000 kcal meal plan]. The portion 300 minutes following the lunch test meal on each study
sizes of the salad dressing and cookies were reduced in the day. The five scales were anchored at the low end with
AI test meal to match the energy and macronutrient con- the most negative feelings (e.g. not at all) and opposing
tent of the C test meal (Table 1). terms at the high end (e.g. extremely high).
The dinner buffet meal on the 3 study days was served
5 hours from the start of the lunch test meal and con- Sample collection and laboratory assessment
tained a variety of foods with pre-identified portion On the 3 assigned study days, participants arrived in the
weight, macronutrient and calorie content. Participants morning to the Nutrition Research Laboratory (NRL) after
were allowed to consume sweet and savory food options a 12-hour overnight fast for a baseline blood draw to meas-
that differed in energy density in an ad libitum manner ure glucose and insulin concentrations. Participants were
to allow for the assessment of postprandial food intake free to engage in their normal morning routines but were
instructed to return to the NRL by noon. The lunch test
Table 1 Percent of daily energy and macronutrient meal (C, AI or AA) was ingested within a 30 minute time
content of the 3 lunch test meals period and additional blood samples were taken at approxi-
Control Avocado Inclusive Avocado Added mately 30, 60, 90, 120 and 180 minutes following the start
Energy, % 35 35 41 of the lunch test meal.
Carbohydrate, % 51 50 49 Venous blood samples were drawn and collected into
Protein, % 14 14 13
vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ)
and centrifuged at 1500 x g at 4°C for 10 min. Serum and
Fat, % 35 36 38
plasma were separated, aliquoted and frozen at −80°C until
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 4 of 9

http://www.nutritionj.com/content/12/1/155

analyzed. Serum glucose was assayed with the glucose- glucose and insulin (Figure 1) and the five VAS questions
oxidase-peroxidase enzymatic assay using kits supplied by (Figure 2), which better represent the reality of the study
Cayman Chemical (Ann Arbor, MI). Serum insulin was data collection and analysis.
assayed using ELISA kits supplied by ALPCO Diagnostics Data are expressed as adjusted mean ± SE unless other-
(Salem, NH). wise noted.

Statistical methods Results

Sample size, power calculations and statistical analysis The 26 participants that completed the study consisted
were performed utilizing SAS version 9.3 (SAS Institute, of 16 women and 10 men with a mean ± SD age 40.8 ±
Cary, NC). All tests were two-sided and a value of P < 11.0 years and BMI 28.1 ± 2.4 kg/m2. Ten participants
0.05 was considered significant. Under good experimen- consumed the 1600 kcal meal plan, nine consumed the
tal laboratory conditions, a sample size of 20 to 25 par-
ticipants has been shown to be adequate to denote a
10% difference in AUC appetite ratings, which is consid-
ered to be a reasonable difference [15]. The target ac-
crual was 28 participants to allow for a 10% dropout
rate, which has been the observed dropout rate for our
prior short-term feeding studies. A mixed model statis-
tical approach was used to compare differences among
the 3 test meals adjusting for study periods as fixed ef-
fects and treating participants as random effects. When
significant findings were observed, Tukey post-hoc test-
ing was performed to further elucidate the differences
between the 3 test meals.
The weighted mean dietary intake of avocado was com-
puted based on the number of participants assigned to
each of the three aforementioned energy levels. The diet-
ary compensation score across the C and AA test meals
was calculated at the individual level using the following
equation [6]: Percent Dietary Compensation = (Intake
without load, C)-(Intake with load, AA)/Energy content of
load x 100. More specifically, the dietary compensation at
dinner was computed at the individual level by subtracting
a subject’s dinner intake after the C lunch test meal minus * *
the same subject’s dinner intake on the day of the AA
lunch test meal, divided by the energy (or macronutrient)
from the avocado consumed.
Two trained research assistants measured the VAS data
to the nearest 0.1 cm and any discrepancy was resolved by
a senior investigator. To compute the area under the
curve (AUC) from zero to x hours (AUC(0-xh)), the mini-
mum value of each subjectively reported VAS scale (in
mm) over time was determined at the individual level and
then the AUC above the minimum value was calculated
using the linear trapezoidal rule. The AUC is reported as
mm x minutes and was constructed by plotting the sub-
jective values between 0 to 100 mm over time (minute)
for each of the five VAS questions.
The VAS was completed and blood samples were sched-
uled to be taken at approximately 30, 60, 90, 120, and Figure 1 Blood glucose and insulin levels after consumption of
180 minutes following the test meal. The exact time for the 3 lunch test meals. Three-hour area under the curve AUC(0-3h)
each individual VAS and blood sample collection was re- based on difference from baseline (time 0) is shown as an insert.
Compared to the AI test meal, the blood insulin was higher in the C
corded and these times were used for the analysis. Curved
and AA test meals (P = 0.04 and P = 0.05, respectively).
lines were generated to show the area under the curve for
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 5 of 9

http://www.nutritionj.com/content/12/1/155

Figure 2 Ratings for the five visual analog scale questions after consumption of the 3 lunch test meals. Five-hour area under the curve
AUC(0-5h) based on difference from baseline (time 0) is shown as an insert. Compared to the C test meal, the AA test meal increased satisfaction
by 23% (P = 0.05) and decreased the desire to eat by 28% (P = 0.04).

2000 kcal meal plan and seven consumed the 2400 kcal and desire to eat (P = 0.05) in the mixed model analysis
meal plan. The weighted mean dietary intake of avocado (Figure 2). Post-hoc testing revealed that compared to the
was 67.5 grams, which contained 112 kcal, 1.3 g protein, C test meal, the AA test meal increased satisfaction by 23%
5.6 g carbohydrate and 10.4 g fat. (P = 0.05) and decreased the desire to eat by 28% (P = 0.04)
for the AUC(0-5h). For the AUC(0-3h), the AA test meal in-
Blood glucose and insulin changes creased satisfaction by 26% (P = 0.02) and decreased the
There were no significant differences between the 3 desire to eat by 40% (P = 0.01) as compared to the C test
lunch test meals for AUC(0-3h) blood glucose (Figure 1). meal (Table 2). Additionally, the AI test meal showed a ten-
Compared to the AI test meal, the AUC(0-3h) for blood dency towards increasing satisfaction by 22% P = 0.07) as
insulin was higher in the C and AA test meals (P = 0.04 compared to the C test meal for the AUC(0-3h). Lastly, the
and P = 0.05, respectively). Difference in blood insulin five measurements of appetite sensation tended to converge
levels between treatments were observed at the 30 mi- 5 hours after the lunch test meal.
nute time point (P = 0.04) as follows: adjusted mean
(95% CI) for C = 54μIU/ml (39, 74); AI = 34μIU/ml (25, Intake at the ad libitum dinner meal and evening snack
47); and, AA 42 μIU/ml (30, 57). Dietary intake at the subsequent ad libitum dinner meal
and evening snack after each of the 3 lunch test meals was
Visual analog scale changes equivalent for total energy, macronutrients and percent
There were significant differences in the AUC(0-5h) for the energy from the macronutrients (Table 3). Compared to
self-reported subjective feelings of satisfaction (P = 0.04) the C test meal, the percent dietary compensation for the
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 6 of 9

http://www.nutritionj.com/content/12/1/155

Table 2 Three-hour area under the curve AUCa(0–3h) based on difference from baseline (time 0) for the five visual analog
scale questions between the 3 lunch test meals
Control Avocado Inclusive Avocado Added
Question Meanb SEc Meanb Differenced P-valuee Meanb Differenced P-valuee P-valuef
How hungry are you? 3105 394 2358 −24% 0.30 2418 −22% 0.36 0.26
How strong is your feeling of fullness? 7249 771 8107 +11% 0.64 8189 +11% 0.58 0.55
How satisfied are you? 6340 726 8149 +22% 0.07 8562 +26% 0.02 0.02
How strong is your desire to eat? 2993 319 2263 −24% 0.16 1798 −40% 0.01 0.01
How much do you think you can eat? 2641 318 2185 −17% 0.45 2031 −23% 0.24 0.25
a
The AUC is reported as mm x minutes and was constructed by plotting the subjective values between 0 to 100 mm over time (minute) for each of the five
VAS questions.
b
Adjusted mean from the mixed model analysis.
c
Common standard error (SE) for all of the adjusted means.
d
Percent difference compared to Control lunch test meal.
e
P-value compared to Control lunch test meal.
f
P-value for diet effect from the mixed model analysis.

AA test meal for energy, protein, carbohydrate and fat was meal as compared to the avocado-free control lunch meal.
66%, 235%, 118% and 36%, respectively. However, the changes in all five measurements of appetite
sensation tended to taper off after 5 hours.
Discussion Energy intake at the subsequent ad libitum dinner
The results of this study suggest that the addition of ~ ½ of meal and evening snack and dietary compensation did
a Hass avocado at a lunch meal can influence post-ingestive not differ between the 3 lunch test meals, which may
satiety over a subsequent 3 hour and 5 hour period in over- have been due to the 5 hour time interval between the
weight and moderately obese adults. Specifically, adding av- lunch test meal and ad libitum dinner meal. De Graaf
ocado to a lunch meal yielded a 23% increase in satisfaction and Hulshof [16] have previously reported that the
(P = 0.05) and a 28% decreased desire to eat (P = 0.04) over weight or amount of food in a preload affects subse-
a subsequent 5 hour period as compared to the avocado- quent appetite and food intake for only up to two hours
free control lunch meal. Also, adding avocado to a lunch after the preload. These findings are consistent with the
meal yielded a 26% increase in satisfaction (P = 0.02) and findings of equivalent energy intake at the subsequent
40% decreased desire to eat (P = 0.01) as compared to the dinner meal and evening snack in the current study, yet
avocado-free control lunch meal over a 3 hour period. inconsistent with changes in two specific measures of
However, an additional 112 kcal was contained in the avo- appetite sensation that we observed at both 3 and
cado, which may have accounted for the observed increased 5 hours for the avocado added test meal. Further, Flint
satisfaction and decreased desire to eat. Further, a 24% de- et al. [15] has reported that an 8-10% difference in the
creased desire to eat (P = 0.16) and 22% increase in satisfac- response magnitude relative to control in food intake or
tion (P = 0.07) was observed over a 3 hour period after satiety score (AUC) is of practical relevance. We found
consumption of the isocaloric avocado inclusive lunch test differences of practical relevance for all five appetite

Table 3 Intake from the dinner meal and evening snack after the 3 lunch test meals
Control (C) Avocado Inclusive (AI) Avocado Added (AA) P-valuec % Dietary Compensationd
Meana SEb Meana Meana Mean SE
Energy (kcal) 1276 82 1193 1194 0.37 66 64
Protein, g (PRO) 42.6 2.8 39.2 39.0 0.27 235 204
Carbohydrate, g (CHO) 134.6 8.6 128.6 126.1 0.37 118 134
Fat, g 64.2 4.6 59.3 60.7 0.47 36 37
PRO, % total energy 13.4 0.3 13.2 13.0 0.36 -
CHO, % total energy 42.8 1.1 43.5 42.6 0.72 -
Fat, % total energy 44.7 1.1 44.3 45.4 0.58 -
a
Adjusted mean based on the mixed model analysis.
b
Common standard error (SE) for all of the adjusted means.
c
P-value for diet effect from the mixed model analysis.
d%Dietary Compensation¼ðIntake without load;CÞ–ðIntake with load;AAÞ  100
Energy content of load .
[Note: The % dietary compensation at dinner was computed at the individual level by subtracting a subject’s dinner intake after the C lunch test meal minus the
same subject’s dinner intake on the day of the AA lunch test meal, divided by the energy (or macronutrient) from the avocado consumed].
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 7 of 9

http://www.nutritionj.com/content/12/1/155

sensation measurements between the C versus the AI the context of increased satisfaction (P = 0.07) and a ten-
and AA interventions ranging between 11-24% and 11- dency existed towards reducing the desire to eat (P =
40%, respectively. However, we did not find a statistically 0.16).
significant difference for hunger, fullness or prospective Insulin and the incretin hormones covary in response
food consumption between the 3 test meals. to elevated postprandial glucose levels [21], which makes
Our overweight participants partially compensated for it challenging to uphold the glucostatic theory proposed
energy (66%) and fat (36%) intake and overcompensated by Mayer [22]. Andersen et al. have observed that post-
for protein (235%) and carbohydrate (118%) at a subse- prandial levels of blood glucose are inversely associated
quent ad libitum dinner meal and evening snack when av- with self-reported appetite and food intake [23], however
ocado (weighted mean energy = 112 kcal) was added to others have shown no association between satiety and
the lunch meal. Thus, the majority of the energy provided blood glucose levels using an intravenous carbohydrate
by the addition of avocado to the diet was offset by dietary infusion [24]. Thus, it is plausible that the incretin hor-
adjustments at the ad libitum dinner meal and evening mones were influenced by the fat and fiber from the
snack. Others have reported that individual daily energy addition of avocado to the AA test meal, which yielded
intake can vary by 20 to 30 percent, and that short-term an increase in satisfaction and a reduction in the desire
dietary manipulations of less than ~400 kcal may not to eat. Although fat delays gastric emptying, some stud-
heavily influence dietary energy compensation [17,18], ies have shown that protein in the diet has the most po-
which may have been one of the reasons for the equivalent tent action on satiety followed by carbohydrate, and fat
subsequent energy intake at the dinner meal and evening the least [25,26]. However, it is important to note that
snack between the 3 study days. studies designed to evaluate the satiety level of fat usu-
There are two potential ways a whole food can be incor- ally add fat to a meal in the form of oil or shortening,
porated into a meal, addition or isocaloric replacement. which increases the energy density of the meal without
Addition is when the food is simply added to a meal, which appreciably altering the volume of the meal. Thus, the
results in an increase in nutrients and total energy, whereas low satiating effect of fat found in some studies may
isocaloric replacement occurs when the food is included have been mediated exclusively by the increase in energy
and other foods are simultaneously decreased or eliminated density.
to compensate for the overall energy content of the meal. It It is also worth noting that the intake at the ad libitum
is worth noting that the AUC(0-3h) for blood glucose in the dinner and evening snack was similar between the AI
current study was equivalent between the 3 lunch test and AA lunch test meals, and that the inclusion of avo-
meals despite the additional mean energy (112 kcal) cado at a meal along with a concurrent reduction in
content and additional ~7 g carbohydrate in the AA lunch other foods containing similar macronutrients favorably
test meal. Avocados contain a unique seven carbon sugar reduced the subsequent energy intake by 83 kcal (6.5%),
(D-manno-heptulose) that does not contribute energy, and and reduced the protein, carbohydrate and fat intake by
some believe it may support blood glucose control and 3.4 g, 6.0 g and 4.9 g, respectively. Avocados are a rich
weight management by reducing glycolysis via hexokinase source of monounsaturated fatty acids, which are prefer-
inhibition [19]. Additionally, 30 minutes after the start of entially oxidized and increase thermogenesis as com-
the lunch test meal the inclusion and addition of avocado pared to polyunsaturated and saturated fatty acids. Thus,
significantly attenuated the rise in blood insulin levels by the inclusion of avocados to a dietary meal pattern may
37% and 22%, respectively (P = 0.04). Avocados are rich in have additional implications in weight management in
antioxidants (e.g. polyphenolic compounds), which others an overweight population.
have shown to be effective in improving insulin sensitivity Although we found a significant reduction in insulin
in an overweight cohort [20]. Hence, including or adding levels and favorable changes in two specific measures of
avocado to a dietary pattern may assist in ameliorating the appetite sensation for the AI and AA lunch test meals, re-
postprandial dysfunction in glucose homeostasis that may spectively, we did not observe any behavioral change in
be present in overweight individuals. dietary intake at the subsequent ad libitum dinner meal
The AUC(0-3h) for blood insulin was lower in the AI test and evening snack between the 3 test meals. However, this
meal compared to both the C and the AA test meals, how- latter null finding should not be over-interpreted as the
ever this biological parameter did not significantly influ- data presented in this study are for 3 separate days (one
ence the five appetite sensation measurements between week apart) and additional dietary energy compensation is
the AI and avocado-free C test meal (P = 0.07 to P = 0.64). plausible over several days and weeks [27].
It is worth noting that the five appetite sensation measure- This study had several strengths and limitations. Our
ments for both the AI and AA test meals went in a favor- controlled “laboratory” type setting had high internal vali-
able and similar direction, and borderline significant dity due to the high degree of sensitivity and control over
findings were found between the AI and C test meal in the dietary intervention and study outcome measures. An
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 8 of 9

http://www.nutritionj.com/content/12/1/155

additional strength is that we analyzed the AUC appetite Acknowledgements

sensation data as opposed to a single time point because The funding for this study was supported by a grant from the Hass Avocado
Board, which had no role in the design and conduct of the study; in the
analysis of individual time points is not physiologically in- collection, analysis, and interpretation of the data; or in the preparation,
dependent and is prone to type 1 errors. A study limita- review or approval of the manuscript.
tion is that we did not measure dietary intake in-between
Author details
the 3 assigned study days. An additional study limitation 1
Department of Nutrition, School of Public Health, Loma Linda University,
is that we provided a wide variety of foods at our ad 24951 N. Circle Dr., Nichol Hall 1102, Loma Linda, CA 92350, USA.
2
libitum dinner buffet meal, which is at variance with the Department of Epidemiology and Biostatistics, Loma Linda University, Loma
Linda, CA, USA.
typical eating pattern of most individuals and is likely to
delay satiation and facilitate increased food intake [28]. Received: 17 August 2013 Accepted: 22 November 2013
Lastly, we may have placed the participants in an atypical Published: 27 November 2013

environment by not providing food to them for 5 hours.

References
1. Giskes K, van Lenthe F, Avendano-Pabon M, Brug J: A systematic review of
Conclusions environmental factors and obesogenic dietary intakes among adults: are
This study showed that the addition of ~ ½ of a fresh we getting closer to understanding obesogenic environments? Obes Rev
Hass avocado to a lunch meal favorably increased satis- 2011, 12:e95–e106.
2. Ouyang H, Yin J, Chen JD: Gastric or intestinal electrical stimulation-induced
faction and reduced the desire to eat over a subsequent increase in gastric volume is correlated with reduced food intake. Scand J
3 hour and 5 hour period in an overweight and moder- Gastroenterol 2006, 41:1261–1266.
ately obese adult population. When avocados were either 3. Rolls BJ: The relationship between dietary energy density and energy
intake. Physiol Behav 2009, 97:609–615.
added or included to a lunch meal, similarities in five 4. Jones PJ, Schoeller DA: Polyunsaturated:saturated ratio of diet fat
measures of appetite sensation were found over a subse- influences energy substrate utilization in the human. Metabolism 1988,
quent 3 hour period. Given that the peak satisfaction ef- 37:145–151.
5. Piers LS, Walker KZ, Stoney RM, Soares MJ, O'Dea K: The influence of the
fect was found to be within 3 hours of the lunch test type of dietary fat on postprandial fat oxidation rates: monounsaturated
meal, subsequent studies should address the offering of (olive oil) vs saturated fat (cream). Int J Obes Relat Metab Disord 2002,
snacks as this may be of importance to overweight and 26:814–821.
6. DiMeglio DP, Mattes RD: Liquid versus solid carbohydrate: effects on food
moderately obese adults that typically consume large intake and body weight. Int J Obes Relat Metab Disord 2000, 24:794–800.
snacks between meals. Therefore, the addition of ~ ½ of 7. Dreher ML, Davenport AJ: Hass avocado composition and potential
an avocado at a specific meal(s) may be a simple dietary health effects. Crit Rev Food Sci Nutr 2013, 53:738–750.
8. Bes-Rastrollo M, van Dam RM, Martinez-Gonzalez MA, Li TY, Sampson LL,
intervention to consider for individuals that consume Hu FB: Prospective study of dietary energy density and weight gain in
large snacks (e.g. excessive energy) between meals. women. Am J Clin Nutr 2008, 88:769–777.
Both the inclusion and addition of avocado to a lunch 9. Fulgoni VL 3rd, Dreher M, Davenport AJ: Avocado consumption is associated
with better diet quality and nutrient intake, and lower metabolic syndrome
meal attenuated the rise in postprandial blood insulin risk in US adults: results from the National Health and Nutrition
levels 30 minutes after the start of the lunch meal as com- Examination Survey (NHANES) 2001–2008. Nutr J 2013, 12:1.
pared to the avocado-free control lunch meal. Addition- 10. Burton-Freeman B: Dietary fiber and energy regulation. J Nutr 2000,
130:272S–275S.
ally, the inclusion of avocado to a lunch meal yielded a 11. Juvonen KR, Purhonen AK, Salmenkallio-Marttila M, Lahteenmaki L,
significant reduction in blood insulin levels over a 3 hour Laaksonen DE, Herzig KH, Uusitupa MI, Poutanen KS, Karhunen LJ: Viscosity
postprandial period. The attenuation in the rise of insulin of oat bran-enriched beverages influences gastrointestinal hormonal
responses in healthy humans. J Nutr 2009, 139:461–466.
in the avocado inclusive intervention is worthy of future 12. Flint A, Gregersen NT, Gluud LL, Moller BK, Raben A, Tetens I, Verdich C,
exploration in persons with insulin resistance and type 2 Astrup A: Associations between postprandial insulin and blood glucose
diabetes mellitus to determine if avocado intake can favor- responses, appetite sensations and energy intake in normal weight and
overweight individuals: a meta-analysis of test meal studies. Br J Nutr
ably influence measures of glucose homeostasis. Lastly, a 2007, 98:17–25.
longer trial would be beneficial to evaluate the effects of 13. Popkin BM, Duffey KJ: Does hunger and satiety drive eating anymore?
daily avocado intake on measures of appetite sensation Increasing eating occasions and decreasing time between eating
occasions in the United States. Am J Clin Nutr 2010, 91:1342–1347.
and weight management in free-living normal weight, 14. Food and Nutrition Board, Institute of Medicine, National Academies: Dietary
overweight and obese adults. Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol,
Protein, and Amino Acids (2002/2005). Dietary Reference Intakes (DRIs): Acceptable
Competing interests Macronutrient Distribution Ranges. Available at: http://www.nap.edu/openbook.
The authors have no conflicts of interest or competing interests. php?record_id=10490&page=1325. Accessed July 10, 2012.
15. Flint A, Raben A, Blundell JE, Astrup A: Reproducibility, power and validity
Authors’ contributions of visual analogue scales in assessment of appetite sensations in single
EH and JS designed the study. EH, MW and JS coordinated the study. MW test meal studies. Int J Obes Relat Metab Disord 2000, 24:38–48.
and EH were responsible for data collection, analysis and quality control. 16. de Graaf C, Hulshof T: Effects of weight and energy content of preloads
MW, EH, JS and KO were involved in the statistical analyses. All authors on subsequent appetite and food intake. Appetite 1996, 26:139–151.
contributed to the interpretation of data. MW wrote the first draft of the 17. Beaton GH, Milner J, Corey P, McGuire V, Cousins M, Stewart E, de Ramos M,
manuscript and all authors critically reviewed and revised the manuscript. JS Hewitt D, Grambsch PV, Kassim N, Little JA: Sources of variance in 24-hour
obtained the funding for the study. All authors read and approved the final dietary recall data: implications for nutrition study design and interpretation.
manuscript. Am J Clin Nutr 1979, 32:2546–2559.
Licensed to MARIA APARECIDA DOS SANTOS FERREIRA - [email protected] - 107.653.668-94

Wien et al. Nutrition Journal 2013, 12:155 Page 9 of 9

http://www.nutritionj.com/content/12/1/155

18. Sukhatme PV, Margen S: Autoregulatory homestatic nature of energy

balance. Am J Clin Nutr 1982, 35:355–365.
19. Roth G, Hayek MSM, Davenport G, Arking R, Bartke A, Bonkowki M,
Ingram D: Mannoheptulose glycolytic inhibitor and novel calorie restriction
mimetic. In Experimental Biology. New Orleans, LA: Abstract # 553.1; 2009.
20. Skulas-Ray AC, Kris-Etherton PM, Teeter DL, Chen CY, Vanden Heuvel JP,
West SG: A high antioxidant spice blend attenuates postprandial insulin
and triglyceride responses and increases some plasma measures of
antioxidant activity in healthy, overweight men. J Nutr 2011, 141:1451–1457.
21. Holt SH, Brand Miller JC, Petocz P: Interrelationships among postprandial
satiety, glucose and insulin responses and changes in subsequent food
intake. Eur J Clin Nutr 1996, 50:788–797.
22. Mayer J: Regulation of energy intake and the body weight: the
glucostatic theory and the lipostatic hypothesis. Ann N Y Acad Sci 1955,
63:15–43.
23. Anderson GH, Catherine NL, Woodend DM, Wolever TM: Inverse
association between the effect of carbohydrates on blood glucose and
subsequent short-term food intake in young men. Am J Clin Nutr 2002,
76:1023–1030.
24. Lavin JH, Wittert G, Sun WM, Horowitz M, Morley JE, Read NW: Appetite
regulation by carbohydrate: role of blood glucose and gastrointestinal
hormones. Am J Physiol 1996, 271:E209–E214.
25. Blundell JE, MacDiarmid JI: Fat as a risk factor for overconsumption:
satiation, satiety, and patterns of eating. J Am Diet Assoc 1997, 97:S63–S69.
26. Rolls BJ, Kim-Harris S, Fischman MW, Foltin RW, Moran TH, Stoner SA:
Satiety after preloads with different amounts of fat and carbohydrate:
implications for obesity. Am J Clin Nutr 1994, 60:476–487.
27. Tarasuk V, Beaton GH: The nature and individuality of within-subject
variation in energy intake. Am J Clin Nutr 1991, 54:464–470.
28. Hetherington MM, Foster R, Newman T, Anderson AS, Norton G:
Understanding variety: tasting different foods delays satiation.
Physiol Behav 2006, 87:263–271.

doi:10.1186/1475-2891-12-155
Cite this article as: Wien et al.: A randomized 3x3 crossover study to
evaluate the effect of Hass avocado intake on post-ingestive satiety,
glucose and insulin levels, and subsequent energy intake in overweight
adults. Nutrition Journal 2013 12:155.

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