John
M de Castro
ABSTRACT by
Sara
Alcohol’s
paying
thing
and
they
eating patterns
of
1-3
Orozco
effects
on
eating
were
humans
to maintain
7-d
ingested,
the time
of ingestion,
their
hol (5). Alcoholism is associated with a myriad of physiological disruptions including liver and pancreatic damage (2, 6, 7), which in turn cause a number ofsecondary changes. These sec-
investigated
diaries
92 adult
of every-
subjective
state
at the time of ingestion, and the number of people present at the time of ingestion. Total intakes, meal sizes, meal compositions, pre- and postmeal intervals, and deprivation and satiety
ondary
changes
may
result
from
either
maldigestion
or malab-
alcohol supplements rather than displaces macronutrient-supplied calories, that alcohol is associated with prolonged meal durations, and that alcohol calories may be unregulated. Other apparent changes in the meal pattern appear to be artifacts of time of day and meal duration. Am J C/in Nutr l990;52:
of the food (5). The observed displacement of nutrients in alcoholics may, thus, be due to these physiological changes rather than to a primary effect ofthe calories supplied by the alcohol. Individuals who consume alcohol but who are not alcoholic appear to add alcohol calories to their calorie intake rather than replace food with alcohol; thus, they consume more total energy than individuals who do not drink (8, 9). Jones et al (10) found that calories derived from alcohol did not replace the calories derived from other nutrients; they were supplemental. Bebb et al (1 1) used food-intake diaries to investigate the nutritional contribution ofalcoholic beverages. They found that on
246-53.
drinking
ratios
were
compared
between
nondrinkers
and
drinkers
and
between meals associated with alcohol ingestion and those without. Univariate and multivariate prediction of meal size and of postmeab interval were also calculated to ascertain alcohol’s
contribution
to the
regulation.
The
results
suggest
that
sorption
days
KEY WORDS carbohydrate,
Meals, protein,
feeding
fat, stomach
pattern,
intermeal
content,
hunger,
ies,
Alcohol is a major component of the typical American diet. It has been estimated that 10 L alcohol are consumed annually per person aged 14 y. This indicates a per capita consumption of 160 kcal/d in the diets ofthe drinking-age population ( 1 ). In many societies alcohol is consumed as both a food and as a drug. As important a constituent of the average diet as it ‘.
is, alcohol’s received
contribution much
to overall
regulation
ofintake
has not
attention.
Calories derived and carbohydrates
from ingested alcohol appear to replace fats as a source of energy in alcoholics (2, 3).
Hillers
(4) found
and
Massey
that
total
energy
intake
ing the
246
effects
of alcohol
on food
intake.
An
mean
is added does
not
arises,
how does
creased
calorie
total
calorie
intake
was
higher
than
on
occurred.
to the diet, as suggested displace
other
nutrients,
by the above then
the
stud-
question
alcohol alter normal regulation, producing inconsumption? Alcohol might be added to normal meals, increasing their size, or it might reduce satiety, increasing the frequency of meals, or both. It might alter the influence of the subjective state on intake (12), of the social facilitation ofeating ( 1 3, 14), or ofthe circadian rhythm ( 1 5). It might change the normal responsiveness to the macronutrients (16) or to the period ofprior deprivation (17, 18). The present study was designed to investigate these possibilities. The effects ofabcohol ingestion on nutrient intake regulation were studied by use ofa database accumulated over several earlier research projects ( 12- 16); new data from ongoing research projects are routinely added.
increased
as alcohol consumption increased, but energy derived from protein, fat, and carbohydrate decreased, as did the average amounts of protein, fat, and carbohydrate consumed in meals. Similarly, Windham et al (1) found that macronutrient intakes were lower in alcoholics than in nondrinkers. These results suggest that the alcoholic is displacing other nutrients as a source of energy. The alcoholic, however, is probably a poor model for studyofalcohol nutrients
and
the
no drinking
Ifalcohob
interval, alcohol
Introduction
days
when
overconsumption
may cause primary malnutrition by displacing other in the diet because ofthe high energy content of alco-
Am J C/in Nuir
Subjects
methods
and
The details (16, l 2
will be briefly
From Georgia Supported
Institute of 87-072 from 3 Address ogy, Georgia Received Accepted
1990;52:246-53.
of the methods
17). They
Printed
State University,
in part by grant
used were published summarized
here.
The
elsewhere study
was
Atlanta.
ROl-DK3988
1-0 1A2 from the National
Diabetes and Digestive and Kidney Diseases and by grant the Georgia State University Research Grant Program. reprint requests to JM de Castro, Department of PsycholState University, University Plaza, Atlanta, GA 30303. June 26, 1989. for publication August 23, 1989.
in USA.
© 1990 American
Society
for Clinical
Nutrition
Downloaded from https://academic.oup.com/ajcn/article-abstract/52/2/246/4651429 by McMaster University Library user on 07 February 2019
Moderate alcohol intake and spontaneous h umans : evidence of unregulated
ALCOHOL conducted
according
Institutes
Health
to the
ethical
guidelines
and the American
of the
Psychological
EFFECTS
male
and
69 female
subjects
recruited
from
a
kg),
and
heightwas
1.68
m(range
1.48-1.94
m).
since the
the
next
interval
subjects and
were
were
given
a small
instructed
to record
(8 X 1 8 cm)
pocket-sized
in as detailed
a manner
as
possible every item that they either ate or drank, the time they ate it, the amount they ate, and how the food was prepared. They were also instructed to record the number of people eating with them and their subjective state ofhunger according to 7-point Likert scale. The subjects recorded for a day and were contacted by the experimenter to review the information, correct any problems, and answer any questions. They then recorded
their
receiving tacted
in the
the subjects
the diary ifany
the
the
diaries
records.
The
names
about
their and
would probably cording period.
any
were
their
intake
phone
entries
Data
in
by phone
were
of two
asked
to pro-
individuals
who
sometime during the rediaries were submitted, and some what
asked to verify difficulty was
the
subject
ate,
the enin no
diary report contradicted in either ofthe food or drink reported.
the
analysis
A computer code
exactly
con-
data
in the diaries.
numbers
in remembering
case was the subject’s nature or the amount
After
and
contacted
the subjects
be eating with them After the completed
them or missing
later
each ofthese individuals was contacted subject’s reported intake. Although countered
days.
reviewed ambiguities
subjects
arose
recording
for 7 consecutive
experimenter
to clarify
questions
Before vide
intake
the diaries
file of
numbers,
was
>
3000
food
used.
The
first
to convert computer
the foods reported codes, indicating
performed
by an experienced
aware
of the experimental
rectly
with
the
subjects.
items, step
which
in the data
had
assigned
analysis
was
in the diaries into the appropriate food types and amounts. This was registered dietician who was unhypotheses and did not interact di-
The
second
step
in the
data
analysis
was the identification ofmeabs and the summation ofthe compositions of the individual items composing the meal. For a reported intake to be classified as an individual meal, it had to contain 50 kcal, or, more stringently, 100 or 200 kcal. The source ofthe calories, (macronutrient, alcohol, solid, or liquid) was not considered. Hence, alcohol ingested without food could be classified as a meal, provided the minimum caloric requirement
was
met.
Each
meal
had
to be separated
premeal
postmeal
(deprivation
interval;
interval; ratio);
the
meal
the
amount
oftime
size divided
postmeab
interval
until
by premeal divided
by
demonstrated,
in
prior
publications
(12-1
7,
19).
Briefly, the reported intake was entered into the model and was estimated to empty from the stomach at a rate proportional to the square root ofthe calorie content ofthe stomach in calories
Methods
The
meal;
meal size (satiety ratio); self ratings of hunger at the beginning ofthe meal; and the estimated premeal and postmeal stomach contents. The calorie content of the stomach was estimated with a computer model that was presented, and that had its efficacy
diary
last
meal;
per minute ach
and
present cluded
(19-22). the
The
proportion
at the beginning in the subsequent
total
food-energy
of each
and
content
macronutrient
at the end
ofthe
stom-
estimated
of each
meal
to be
was
in-
analyses.
Summaries of the intake characteristics of each individual subject were prepared by calculating means ofthe meal characteristics across all meals, meals eaten with alcohol, and meals eaten without alcohol. Pearson product-moment correlations were calculated for each subject between the amount of alcohol consumed in the meal and the amount of other nutrients in the meal, the postmeal interval, and the postmeal interval plus meal duration. They were calculated separately for all meals, for meals that were reported to have been accompanied by alcohol, and for meals eaten without alcohol. Correlations were also calculated between the number of people present and the meal characteristics and intervals. Correlations were calculated by use ofall
intermeal
intervals
except
the overnight
fast.
Mul-
tiple-linear-regression analyses were also performed on these same data individually for each subject. The total calorie content ofthe meal, the alcohol content ofthe meal, the nonalcohol calorie content ofthe meal, and the postmeal interval were used as dependent, predicted variables whereas the number of people present, the self-ratings, the time of day, the stomach contents, premeal interval, and the alcohol content ofthe meal were used as independent, predictor variables. (See ref 13 for the rationale for the choice ofthese factors.) Group means and SEMs were then calculated by using the meal characteristics, correlations, multiple-regression (standardized)
coefficients,
intercepts,
and
multiple
correlations
that had been calculated for each subject individually (23). Before further analysis the correlation coefficients were transformed to z scores (24). The mean correlations and coefficients were then compared with 0 by use of a t test. Within-group mean correlations, fi coefficients, or intercepts were compared with a correlated-groups t test.
Results
in time
from the preceding meal and the ingestive behaviors that folbowed the meal by 1S mm. More stringent definitions of 45 and 90 mm were also employed. Five different definitions of a meal were used, combining these minimum criteria: 1S mm and SO kcal, 45 mm and SO kcal, 45 mm and 100 kcal, 45 mm and 200 kcal, and 90 mm and 50 kcal.
Of the 92 subjects 32 did not ingest any alcohol during the recording period; they were designated the no-alcohol group. The remaining 60 subjects all reported alcohol ingestion and were further subdivided by a median split into low- and moderate-alcohol-consumption groups. The groups thus constituted were comparable in height (1 .67 ± 0.02, 1.65 ± 0.02, and I .69
Downloaded from https://academic.oup.com/ajcn/article-abstract/52/2/246/4651429 by McMaster University Library user on 07 February 2019
newspaper ad and by word of mouth were paid $30 to participate. They also received a detailed nutrition analysis based on their food intake for the 7-d reporting period. They were told that this was a study of nutrient intakes in humans. Average age was 3 1.6 y(range 18-54 y), weight was 61.8 kg (range 45.5102.3
247
EATING
Meals were characterized by their total calorie content; the proportion ofthe recommended dietary allowance for food energy ingested (18); alcohol, carbohydrate, fat, and protein content; the total nonalcohol calorie content; the amount of time
National
Association.
Subjects Twenty-three
ON
248
DE
CASTRO
AND
OROZCO
solely
to the alcohol
calories
consumed;
macronutrient
were equivalent. Hence, even when viewed cohol calories appear to add to total intake place
other
Meal
characteristics were
performed meal
cant qualitative ent definitions. the
on meals
Low
Mod#{149},-ot#{149} Intoc#{149} day.
olcoho)
WI thout without
descriptive
and
inferential
olcoho)
with
469
FIG I . Total
daily caloric intakes (± SEM) ofcarbohydrate (solid), fat (cross-hatched), protein (hatched), and alcohol (unshaded) for the subjects who spontaneously ingested no, low, or moderate amounts of alcohol (left three bars), and, for the drinkers, the average caloric intakes for days during which alcohol was ingested and days during which it was not (right two bars).
weight (62.7 index (12.85
± 2. 1 , 59.0 ± 2.3, and ± 0.10, 12.97 ± 0.10,
63.9 and
12.93 ± 0. 10), sex (22%, 23%, and 30% for males), and age (29.2 1.6, 3 1.7 ± 1.5, and 34.2 ± 1 .6 y) for the no-, low- and moder-
±
ate-alcohol
tween
groups,
respectively.
the no-alcohol
tistically
The
mean
(p < 0.05). calorie intakes
daily
reported are presented
erate-alcohol
groups
consisting
of 859,
for
ingested
respectively.
gested
significantly
more
cantly
greater
slightly
group.
greater
their
total
amount
was 5th-
for
kcal
2026,
and
2174
carbohydrate;
calories
(p
0.05)
M de Castro
ABSTRACT by
Sara
Alcohol’s
paying
thing
and
they
eating patterns
of
1-3
Orozco
effects
on
eating
were
humans
to maintain
7-d
ingested,
the time
of ingestion,
their
hol (5). Alcoholism is associated with a myriad of physiological disruptions including liver and pancreatic damage (2, 6, 7), which in turn cause a number ofsecondary changes. These sec-
investigated
diaries
92 adult
of every-
subjective
state
at the time of ingestion, and the number of people present at the time of ingestion. Total intakes, meal sizes, meal compositions, pre- and postmeal intervals, and deprivation and satiety
ondary
changes
may
result
from
either
maldigestion
or malab-
alcohol supplements rather than displaces macronutrient-supplied calories, that alcohol is associated with prolonged meal durations, and that alcohol calories may be unregulated. Other apparent changes in the meal pattern appear to be artifacts of time of day and meal duration. Am J C/in Nutr l990;52:
of the food (5). The observed displacement of nutrients in alcoholics may, thus, be due to these physiological changes rather than to a primary effect ofthe calories supplied by the alcohol. Individuals who consume alcohol but who are not alcoholic appear to add alcohol calories to their calorie intake rather than replace food with alcohol; thus, they consume more total energy than individuals who do not drink (8, 9). Jones et al (10) found that calories derived from alcohol did not replace the calories derived from other nutrients; they were supplemental. Bebb et al (1 1) used food-intake diaries to investigate the nutritional contribution ofalcoholic beverages. They found that on
246-53.
drinking
ratios
were
compared
between
nondrinkers
and
drinkers
and
between meals associated with alcohol ingestion and those without. Univariate and multivariate prediction of meal size and of postmeab interval were also calculated to ascertain alcohol’s
contribution
to the
regulation.
The
results
suggest
that
sorption
days
KEY WORDS carbohydrate,
Meals, protein,
feeding
fat, stomach
pattern,
intermeal
content,
hunger,
ies,
Alcohol is a major component of the typical American diet. It has been estimated that 10 L alcohol are consumed annually per person aged 14 y. This indicates a per capita consumption of 160 kcal/d in the diets ofthe drinking-age population ( 1 ). In many societies alcohol is consumed as both a food and as a drug. As important a constituent of the average diet as it ‘.
is, alcohol’s received
contribution much
to overall
regulation
ofintake
has not
attention.
Calories derived and carbohydrates
from ingested alcohol appear to replace fats as a source of energy in alcoholics (2, 3).
Hillers
(4) found
and
Massey
that
total
energy
intake
ing the
246
effects
of alcohol
on food
intake.
An
mean
is added does
not
arises,
how does
creased
calorie
total
calorie
intake
was
higher
than
on
occurred.
to the diet, as suggested displace
other
nutrients,
by the above then
the
stud-
question
alcohol alter normal regulation, producing inconsumption? Alcohol might be added to normal meals, increasing their size, or it might reduce satiety, increasing the frequency of meals, or both. It might alter the influence of the subjective state on intake (12), of the social facilitation ofeating ( 1 3, 14), or ofthe circadian rhythm ( 1 5). It might change the normal responsiveness to the macronutrients (16) or to the period ofprior deprivation (17, 18). The present study was designed to investigate these possibilities. The effects ofabcohol ingestion on nutrient intake regulation were studied by use ofa database accumulated over several earlier research projects ( 12- 16); new data from ongoing research projects are routinely added.
increased
as alcohol consumption increased, but energy derived from protein, fat, and carbohydrate decreased, as did the average amounts of protein, fat, and carbohydrate consumed in meals. Similarly, Windham et al (1) found that macronutrient intakes were lower in alcoholics than in nondrinkers. These results suggest that the alcoholic is displacing other nutrients as a source of energy. The alcoholic, however, is probably a poor model for studyofalcohol nutrients
and
the
no drinking
Ifalcohob
interval, alcohol
Introduction
days
when
overconsumption
may cause primary malnutrition by displacing other in the diet because ofthe high energy content of alco-
Am J C/in Nuir
Subjects
methods
and
The details (16, l 2
will be briefly
From Georgia Supported
Institute of 87-072 from 3 Address ogy, Georgia Received Accepted
1990;52:246-53.
of the methods
17). They
Printed
State University,
in part by grant
used were published summarized
here.
The
elsewhere study
was
Atlanta.
ROl-DK3988
1-0 1A2 from the National
Diabetes and Digestive and Kidney Diseases and by grant the Georgia State University Research Grant Program. reprint requests to JM de Castro, Department of PsycholState University, University Plaza, Atlanta, GA 30303. June 26, 1989. for publication August 23, 1989.
in USA.
© 1990 American
Society
for Clinical
Nutrition
Downloaded from https://academic.oup.com/ajcn/article-abstract/52/2/246/4651429 by McMaster University Library user on 07 February 2019
Moderate alcohol intake and spontaneous h umans : evidence of unregulated
ALCOHOL conducted
according
Institutes
Health
to the
ethical
guidelines
and the American
of the
Psychological
EFFECTS
male
and
69 female
subjects
recruited
from
a
kg),
and
heightwas
1.68
m(range
1.48-1.94
m).
since the
the
next
interval
subjects and
were
were
given
a small
instructed
to record
(8 X 1 8 cm)
pocket-sized
in as detailed
a manner
as
possible every item that they either ate or drank, the time they ate it, the amount they ate, and how the food was prepared. They were also instructed to record the number of people eating with them and their subjective state ofhunger according to 7-point Likert scale. The subjects recorded for a day and were contacted by the experimenter to review the information, correct any problems, and answer any questions. They then recorded
their
receiving tacted
in the
the subjects
the diary ifany
the
the
diaries
records.
The
names
about
their and
would probably cording period.
any
were
their
intake
phone
entries
Data
in
by phone
were
of two
asked
to pro-
individuals
who
sometime during the rediaries were submitted, and some what
asked to verify difficulty was
the
subject
ate,
the enin no
diary report contradicted in either ofthe food or drink reported.
the
analysis
A computer code
exactly
con-
data
in the diaries.
numbers
in remembering
case was the subject’s nature or the amount
After
and
contacted
the subjects
be eating with them After the completed
them or missing
later
each ofthese individuals was contacted subject’s reported intake. Although countered
days.
reviewed ambiguities
subjects
arose
recording
for 7 consecutive
experimenter
to clarify
questions
Before vide
intake
the diaries
file of
numbers,
was
>
3000
food
used.
The
first
to convert computer
the foods reported codes, indicating
performed
by an experienced
aware
of the experimental
rectly
with
the
subjects.
items, step
which
in the data
had
assigned
analysis
was
in the diaries into the appropriate food types and amounts. This was registered dietician who was unhypotheses and did not interact di-
The
second
step
in the
data
analysis
was the identification ofmeabs and the summation ofthe compositions of the individual items composing the meal. For a reported intake to be classified as an individual meal, it had to contain 50 kcal, or, more stringently, 100 or 200 kcal. The source ofthe calories, (macronutrient, alcohol, solid, or liquid) was not considered. Hence, alcohol ingested without food could be classified as a meal, provided the minimum caloric requirement
was
met.
Each
meal
had
to be separated
premeal
postmeal
(deprivation
interval;
interval; ratio);
the
meal
the
amount
oftime
size divided
postmeab
interval
until
by premeal divided
by
demonstrated,
in
prior
publications
(12-1
7,
19).
Briefly, the reported intake was entered into the model and was estimated to empty from the stomach at a rate proportional to the square root ofthe calorie content ofthe stomach in calories
Methods
The
meal;
meal size (satiety ratio); self ratings of hunger at the beginning ofthe meal; and the estimated premeal and postmeal stomach contents. The calorie content of the stomach was estimated with a computer model that was presented, and that had its efficacy
diary
last
meal;
per minute ach
and
present cluded
(19-22). the
The
proportion
at the beginning in the subsequent
total
food-energy
of each
and
content
macronutrient
at the end
ofthe
stom-
estimated
of each
meal
to be
was
in-
analyses.
Summaries of the intake characteristics of each individual subject were prepared by calculating means ofthe meal characteristics across all meals, meals eaten with alcohol, and meals eaten without alcohol. Pearson product-moment correlations were calculated for each subject between the amount of alcohol consumed in the meal and the amount of other nutrients in the meal, the postmeal interval, and the postmeal interval plus meal duration. They were calculated separately for all meals, for meals that were reported to have been accompanied by alcohol, and for meals eaten without alcohol. Correlations were also calculated between the number of people present and the meal characteristics and intervals. Correlations were calculated by use ofall
intermeal
intervals
except
the overnight
fast.
Mul-
tiple-linear-regression analyses were also performed on these same data individually for each subject. The total calorie content ofthe meal, the alcohol content ofthe meal, the nonalcohol calorie content ofthe meal, and the postmeal interval were used as dependent, predicted variables whereas the number of people present, the self-ratings, the time of day, the stomach contents, premeal interval, and the alcohol content ofthe meal were used as independent, predictor variables. (See ref 13 for the rationale for the choice ofthese factors.) Group means and SEMs were then calculated by using the meal characteristics, correlations, multiple-regression (standardized)
coefficients,
intercepts,
and
multiple
correlations
that had been calculated for each subject individually (23). Before further analysis the correlation coefficients were transformed to z scores (24). The mean correlations and coefficients were then compared with 0 by use of a t test. Within-group mean correlations, fi coefficients, or intercepts were compared with a correlated-groups t test.
Results
in time
from the preceding meal and the ingestive behaviors that folbowed the meal by 1S mm. More stringent definitions of 45 and 90 mm were also employed. Five different definitions of a meal were used, combining these minimum criteria: 1S mm and SO kcal, 45 mm and SO kcal, 45 mm and 100 kcal, 45 mm and 200 kcal, and 90 mm and 50 kcal.
Of the 92 subjects 32 did not ingest any alcohol during the recording period; they were designated the no-alcohol group. The remaining 60 subjects all reported alcohol ingestion and were further subdivided by a median split into low- and moderate-alcohol-consumption groups. The groups thus constituted were comparable in height (1 .67 ± 0.02, 1.65 ± 0.02, and I .69
Downloaded from https://academic.oup.com/ajcn/article-abstract/52/2/246/4651429 by McMaster University Library user on 07 February 2019
newspaper ad and by word of mouth were paid $30 to participate. They also received a detailed nutrition analysis based on their food intake for the 7-d reporting period. They were told that this was a study of nutrient intakes in humans. Average age was 3 1.6 y(range 18-54 y), weight was 61.8 kg (range 45.5102.3
247
EATING
Meals were characterized by their total calorie content; the proportion ofthe recommended dietary allowance for food energy ingested (18); alcohol, carbohydrate, fat, and protein content; the total nonalcohol calorie content; the amount of time
National
Association.
Subjects Twenty-three
ON
248
DE
CASTRO
AND
OROZCO
solely
to the alcohol
calories
consumed;
macronutrient
were equivalent. Hence, even when viewed cohol calories appear to add to total intake place
other
Meal
characteristics were
performed meal
cant qualitative ent definitions. the
on meals
Low
Mod#{149},-ot#{149} Intoc#{149} day.
olcoho)
WI thout without
descriptive
and
inferential
olcoho)
with
469
FIG I . Total
daily caloric intakes (± SEM) ofcarbohydrate (solid), fat (cross-hatched), protein (hatched), and alcohol (unshaded) for the subjects who spontaneously ingested no, low, or moderate amounts of alcohol (left three bars), and, for the drinkers, the average caloric intakes for days during which alcohol was ingested and days during which it was not (right two bars).
weight (62.7 index (12.85
± 2. 1 , 59.0 ± 2.3, and ± 0.10, 12.97 ± 0.10,
63.9 and
12.93 ± 0. 10), sex (22%, 23%, and 30% for males), and age (29.2 1.6, 3 1.7 ± 1.5, and 34.2 ± 1 .6 y) for the no-, low- and moder-
±
ate-alcohol
tween
groups,
respectively.
the no-alcohol
tistically
The
mean
(p < 0.05). calorie intakes
daily
reported are presented
erate-alcohol
groups
consisting
of 859,
for
ingested
respectively.
gested
significantly
more
cantly
greater
slightly
group.
greater
their
total
amount
was 5th-
for
kcal
2026,
and
2174
carbohydrate;
calories
(p
0.05)
