The Effects of Low-Carbohydrate versus Conventional Weight
Loss Diets in Severely Obese Adults: One-Year Follow-up of a
Linda Stern, MD; Nayyar Iqbal, MD; Prakash Seshadri, MD; Kathryn L. Chicano, CRNP; Denise A. Daily, RD; Joyce McGrory, CRNP;
Monica Williams, BS; Edward J. Gracely, PhD; and Frederick F. Samaha, MD
Background: A previous paper reported the 6-month compari-
1.0 kg]; P
؍ 0.20). For persons on the low-carbohydrate diet,
son of weight loss and metabolic changes in obese adults ran-
triglyceride levels decreased more (P
؍ 0.044) and high-density
domly assigned to either a low-carbohydrate diet or a conven-
lipoprotein cholesterol levels decreased less (P
؍ 0.025). As seen
tional weight loss diet.
in the small group of persons with diabetes (n
؍ 54) and after
adjustment for covariates, hemoglobin A
levels improved more
To review the 1-year outcomes between these diets.
for persons on the low-carbohydrate diet. These more favorable
Design: Randomized trial.
metabolic responses to a low-carbohydrate diet remained signifi-
cant after adjustment for weight loss differences. Changes in other
Setting: Philadelphia Veterans Affairs Medical Center.
lipids or insulin sensitivity did not differ between groups.
Participants: 132 obese adults with a body mass index of 35
Limitations: These findings are limited by a high dropout rate
kg/m2 or greater; 83% had diabetes or the metabolic syndrome.
(34%) and by suboptimal dietary adherence of the enrolled per-
Intervention: Participants received counseling to either restrict
carbohydrate intake to <30 g per day (low-carbohydrate diet) or
Conclusion: Participants on a low-carbohydrate diet had more
to restrict caloric intake by 500 calories per day with <30% of
favorable overall outcomes at 1 year than did those on a conven-
calories from fat (conventional diet).
tional diet. Weight loss was similar between groups, but effects
on atherogenic dyslipidemia and glycemic control were still more
Measurements: Changes in weight, lipid levels, glycemic con-
favorable with a low-carbohydrate diet after adjustment for differ-
trol, and insulin sensitivity.
ences in weight loss.
Results: By 1 year, mean (±SD) weight change for persons on the
Ann Intern Med.
low-carbohydrate diet was
؊5.1 ± 8.7 kg compared with
For author affiliations, see end of text.
؊3.1 ± 8.4 kg for persons on the conventional diet. Differences
See related article on pp 769-777 and editorial comment on pp 836-
between groups were not significant (
؊1.9 kg [95% CI,
he prevalence of obesity and its associated metabolic results would be important, given the high-risk nature of
abnormalities has increased markedly over the past 2
our study sample, but that long-term outcomes would pro-
decades (1, 2). Although guidelines to follow a high–com-
vide more information about the sustainability of any diet-
plex carbohydrate, low-fat, energy-deficient diet to achieve
related outcomes. We now report our findings 1 year after
weight loss are generally accepted (3), considerable public
randomization to a low-carbohydrate diet versus a low-fat
interest has focused on low-carbohydrate diets (4). We re-
weight loss diet (conventional diet) in severely obese adults
cently reported that persons with severe obesity lost more
with a high prevalence of diabetes or the metabolic syn-
weight and had greater improvements in triglyceride levels,
insulin sensitivity, and glycemic control after 6 months of alow-carbohydrate diet as compared with a conventionalweight loss diet based on calorie and fat restriction (5).
However, these findings were preliminary because of the
short duration of that study (6). A simultaneously pub-
The study design has been previously described (5).
lished study by Foster and colleagues suggested that per-
Participants were recruited from the outpatient practices of
sons on a low-carbohydrate diet tended to regain weight by
the Philadelphia Veterans Affairs Medical Center and in-
1 year (7). These findings were limited, however, because
cluded persons 18 years of age and older with a body mass
few participants completed the study and because the study
index (BMI) of 35 kg/m2 or greater. The exclusion criteria
used a self-help approach, which is less effective than direct
were a serum creatinine level greater than 133 mol/L
counseling for maintaining weight loss (8). Foster and col-
(Ͼ1.5 mg/dL), hepatic disease, severe life-limiting medical
leagues also excluded persons with diabetes, which is highly
illness, inability to self-monitor glucose levels, or active use
of a weight loss program or weight loss medication. Be-
During the development of this study, we decided to
tween May 2001 and November 2001, 132 persons were
analyze and report preliminary results at 6 months and
randomly assigned to either a low-carbohydrate diet (n
final results at 1 year. We thought that the short-term
64) or a conventional diet (n
ϭ 68). The Institutional Re-
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10
view Committee at the Philadelphia Veterans Affairs Med-ical Center approved the study, and all participants pro-
In 2003, the authors reported that severely obese adults
lost more weight and had better serum lipid patterns after
Diet groups met in weekly counseling sessions for 4
6 months of a low-carbohydrate diet rather than a con-ventional low-fat diet.
weeks, followed by 11 monthly sessions. Participants onthe low-carbohydrate diet were instructed only to reduce
carbohydrate intake to less than 30 g per day. Participants
After 1 year, these same patients still had more favorable
on the conventional diet were instructed to reduce caloric
triglyceride and high-density lipoprotein cholesterol levels
intake by 500 calories per day, with less than 30% of cal-
on the low-carbohydrate diet than on the conventional
ories derived from fat, in accordance with the National
diet. However, weight loss and the other metabolic pa-
Heart, Lung, and Blood Institute guidelines (3).
rameters were similar in the 2 diet groups.
We collected data, including weight (single calibrated
The effect of the modest improvements in high-density
scale, SR Instruments, Inc., Tonawanda, New York), med-
lipoprotein cholesterol and triglyceride levels on the devel-
ical history (self-reported), and blood pressure, at baseline,
opment of diabetes and cardiovascular disease is un-
6 months, and 1 year. Fasting blood specimens were ob-
tained for glucose, hemoglobin A1c, and serum lipid levels(Synchron LX20, Beckman Coulter, Inc., Fullerton, Cali-
fornia). Low-density lipoprotein (LDL) cholesterol levelwas calculated by using the Friedewald formula (9). Wedefined the presence of diabetes by a historical fasting
rithm generated from a random set of numbers that was
blood glucose level greater than 6.94 mmol/L (Ͼ125 mg/
constructed and held in a separate center and concealed
dL) or use of antidiabetic medications. The metabolic syn-
from those enrolling persons during randomization. We
drome was considered present if a participant had 3 or
used stratified randomization, with blocking within strata,
more of the following (10): central obesity, fasting blood
to ensure assignment of approximately equal numbers of
glucose level of 6.11 mmol/L (110 mg/dL) or greater, fast-
women, diabetic persons, and severely obese persons
ing triglyceride level of 1.70 mmol/L (150 mg/dL) or
(BMI Ն 40 kg/m2) to each study group.
greater, high-density lipoprotein (HDL) cholesterol level
Changes in weight, dietary intake, and metabolic data
less than 1.04 mmol/L (Ͻ40 mg/dL) for men or less than
were compared between the 2 diets by random-coefficient
1.30 mmol/L (Ͻ50 mg/dL) for women, blood pressure of
analysis (11). This type of analysis was selected to allow for
130/85 mm Hg or greater, or antihypertensive therapy.
a variable number of observations for participants and to
We assumed that all participants had central obesity be-
take into account that the repeated observations of the
cause of the uniform severity of their obesity (BMI range,
outcome variables over time for individuals were corre-
35.0 to 79.4 kg/m2). Serum insulin was measured by ra-
lated. The random-coefficient analysis model takes these
dioimmunoassay (Laboratory Corporation of America
correlations into account by allowing the intercept to vary
Holdings [LabCorp], Burlington, North Carolina]). Insu-
randomly among persons. We used a restricted maximum
lin resistance in nondiabetic persons was estimated by the
likelihood analysis, which assumed that changes were dis-
quantitative insulin sensitivity check (QUICK) index:
tributed according to a bivariate normal distribution and
1/[(log (fasting insulin (U/mL)) ϩ (log fasting glu-
that data were missing at random. The outcome variables
were changes from baseline in weight, dietary macronutri-
ent consumption, and metabolic measurements. For all of
Our primary end point was total weight loss at 1 year.
these analyses, the covariates included an indicator variable
Secondary analyses included the change from baseline in
for time (6 months and 1 year), diet group, and a diet
serum lipid levels, insulin sensitivity, and glycemic control.
group by time interaction term. This diet group by time
We estimated that we would need 100 persons (50 per
interaction term was kept in the model, regardless of its
group), assuming a 2-sided type I error of 5%, for the
statistical significance (P
ϭ 0.063 for the weight loss anal-
study to have 80% power to detect a 5-kg greater mean
ysis). Separate analyses to adjust for baseline differences
weight loss in the low-carbohydrate group than in the con-
between diet groups were also made by entering the fol-
ventional diet group. These calculations were based on an
lowing covariates to each of these models: age; race (white
anticipated maximum weight loss by 6 months, with
or African American); sex; baseline BMI; baseline caloric
weight stabilization in both diet groups between 6 months
intake; and the presence or absence of hypertension, use of
and 1 year. To compensate for an anticipated dropout rate
lipid-lowering therapy, diabetes, active smoking, and sleep
of 25%, we set our enrollment target at 135 persons. Ran-
apnea (12). All variables were assessed for normality before
domization was performed by using a pre-established algo-
entry into the analyses. Triglyceride, insulin, and glucose
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10 779
Table 1. Baseline Characteristics of Study Participants*
in the Study
Dropped Out of
in the Study
Dropped Out of
Metabolic syndrome (without diabetes), %
* Values expressed with a plus/minus sign are the mean Ϯ SD. Participants were considered to have hyperlipidemia if they reported a total cholesterol level greater than 5.18
mmol/L (Ͼ200 mg/dL) or were actively using lipid-lowering therapy. P
values were determined by chi-square analysis for categorical variables, by the unpaired t
continuous variables, and by logistic regression for interaction terms. BMI ϭ body mass index.
ϭ 0.001 for younger age predicting a greater likelihood of dropping out of the study.
Ͼ 0.2 for baseline differences in race between diet groups; P
ϭ 0.162 for an interaction between diet and race on the number of persons who dropped out at 1 year.
ϭ 0.120 for an interaction between diet and sex on the number of persons who dropped out at 1 year.
Includes 2 persons who developed diabetes during the first few weeks after study enrollment.
ϭ 0.082 for a difference in the prevalence of hypertension between diet groups.
ϭ 0.053 for the ability of the presence of sleep apnea to predict a greater likelihood of remaining in the study.
levels were skewed and thus were log-transformed before
(total, 39 persons at a mean [ϮSD] of 13.5 Ϯ 3.2
the analyses. Baseline differences between diet groups were
months). Thus, we had 6-month weights on 118 of 132
compared by chi-square analysis for dichotomous variables
persons (89%) and 1-year weights on 126 of 132 persons
and by the unpaired t
-test for continuous variables. All P
(96%). Of the 18 persons who missed the 6-month visit
values are 2-sided, and a P
value of 0.05 was considered
but returned for the 1-year visit (6 in the low-carbohydrate
statistically significant. Analyses were performed with SPSS
group and 12 in the conventional diet group), all but 2 had
statistical software, version 11.1 (SPSS, Inc., Chicago, Illi-
6-month weights retrieved from medical records. Of the 6
persons for whom no 1-year weights were available, 2 were
in the low-carbohydrate group and 4 in the conventional
Of the 132 enrolled persons, follow-up was done at 6
diet group. The weights retrieved from medical records
months for 79 persons and at 1 year for 87 persons. For
were obtained on scales that were different from those used
measurements at 6 months, we retrieved weights on an
for the study and were probably obtained in a nonuniform
additional 16 persons on the low-carbohydrate diet and 23
persons on the conventional diet (total, 39 persons at a
We used several approaches to handle the 45 partici-
mean [ϮSD] of 6.6 Ϯ 1.2 months). For measurements at
pants with missing data for diet recall and metabolic mea-
1 year, we retrieved weights on 18 persons on the low-
surements. For the primary analysis by random-coefficient
carbohydrate diet and 21 persons on the conventional diet
analysis, we assumed data were missing at random. To
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10
verify this assumption, we performed sensitivity analyses
those who dropped out of the study and those who com-
based on comparisons of baseline characteristics and weight
pleted the study were not significant.
loss differences between those who dropped out and those
Dietary Intake Assessment
who completed the study. We also performed 2 additional
shows the dietary recall data. Caloric intake
sensitivity analyses: The first analysis included only persons
decreased more by 1 year in the low-carbohydrate group
who completed the study, and the second analysis included
than in the conventional diet group, although the differ-
all persons, with the baseline data carried forward for those
ence between diet groups was not statistically significant.
The low-carbohydrate group reduced carbohydrate intake
Role of the Funding Source
by 52%, reduced fiber intake by 42%, increased total fat
The funding source had no role in the design, con-
intake by 31%, increased dietary cholesterol intake by
duct, or reporting of the study or in the decision to submit
32%, and reduced sodium intake by 21% relative to base-
line. However, only the reductions in carbohydrate intakeand sodium intake were greater than observed in the con-
Participants were well matched between diet groups
Participants on the low-carbohydrate diet maintained
regarding baseline characteristics, although the low-carbo-
most of their 6-month weight loss, whereas those on a
hydrate group had more hypertensive and white persons
conventional diet continued to lose weight throughout the
). Both groups had a high prevalence of diabetes
year. The final 1-year weight change (mean Ϯ SD) was
or the metabolic syndrome (Table 1
). Twenty persons on
Ϫ5.1 Ϯ 8.7 kg in the low-carbohydrate group and
the low-carbohydrate diet and 25 on the conventional diet
Ϫ3.1 Ϯ 8.4 kg in the conventional diet group (Figure
dropped out by 1 year. These persons were younger and
The difference in weight loss between the 2 diet groups was
had a lower prevalence of sleep apnea but were not other-
not significant (Ϫ2.0 kg [CI, Ϫ4.9 kg to 1.0 kg]; P
wise significantly different from those who completed the
0.195 before and P
Ͼ 0.2 after adjustment for baseline
study (Table 1
). Differences in baseline lipid values (P
variables). The difference in weight loss between the 2 diet
0.2 for all comparisons), diet composition (P
Ն 0.149 for
groups between 6 months and 1 year was not statistically
all comparisons), glycemic control indices (P
Ն 0.158 for
ϭ 0.063). Persons on the low-carbohydrate
all comparisons), and insulin sensitivity (P
Ͼ 0.2) between
diet who dropped out lost less weight than those who com-
Table 2. Changes in Dietary Composition between Baseline and 1 Year for the 2 Diets*
1 Year (n
Mean Difference (95% CI)†
Saturated fat, g
Dietary cholesterol, g
Dietary sodium, mg
* The values for dietary macronutrient data are given as the mean (ϮSD) g per day and are based on 24-hour dietary recall from the 87 persons who completed the study.
† The mean difference in 1-year change and 95% CIs are for the low-carbohydrate group relative to the conventional diet group and are obtained by random-coefficientanalysis.
‡ The P
values are for comparison of the change from baseline to 1 year between diet groups by random-coefficient analysis. All participants are included and the values arenot adjusted for baseline variables.
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10 781
Table 3. Change in Serum Lipids, Glycemic Indices, Creatinine Levels, and Uric Acid Levels at 1 Year*
Triglyceride level, mmol/L
Cholesterol level, mmol/L
HDL cholesterol level, mmol/L
LDL cholesterol level, mmol/L
Glucose level for persons
without diabetes, mmol/L
Glucose level for persons with
diabetes, mmol/L (mg/dL)
1c level for persons with
diabetes, % (n
Insulin level for persons
without diabetes, pmol/L
Insulin level in persons with
Insulin sensitivity in persons
Serum creatinine level,
Blood urea nitrogen level,
Uric acid level, mmol/L
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10
Blood pressure, mm Hg
* Values are given as the mean (ϮSD) for the 87 persons who completed the study. HBA ϭ hemoglobin A ; HDL ϭ high-density lipoprotein; LDL ϭ low-density
† The mean difference in 1-year change and 95% CIs are for the low-carbohydrate diet group relative to the conventional diet group and are obtained by random-coefficientanalysis.
‡ The P
values are for comparison of the change from baseline to 1 year between diet groups by random-coefficient analysis.
§ Adjusted P
values are obtained by random-coefficient analysis; the included covariates are age, race, sex, baseline body mass index, baseline caloric intake, and the presenceor absence of hypertension, lipid-lowering therapy use, diabetes, active smoking, and sleep apnea.
LDL cholesterol was not calculated in 2 persons in the low-carbohydrate diet group because of triglyceride levels above 4.52 mmol/L (400 mg/dL).
¶Determined by the quantitative insulin sensitivity check index: 1/[(fasting insulin (U/mL)) ϩ (log fasting glucose (mg/dL))].
pleted the study (change, Ϫ0.2 Ϯ 7.6 kg vs. Ϫ7.3 Ϯ 8.3
Glycemic Control and Insulin Sensitivity
kg, respectively; mean difference, Ϫ7.1 kg [CI, Ϫ11.6 kg
The difference in the response of glucose and insulin
to Ϫ2.8 kg]; P
ϭ 0.003). In contrast, weight loss was not
sensitivity between diet groups by 1 year was not signifi-
significantly different for those on the conventional diet,
cant (Table 3
). Despite this, the hemoglobin A1c level in
whether they dropped out or completed the study (change,
the small group of persons with diabetes (n
ϭ 54) de-
Ϫ2.2 Ϯ 9.5 kg vs. Ϫ3.7 Ϯ 7.7, respectively; mean differ-
creased more in the low-carbohydrate group, after adjust-
ence, Ϫ1.5 kg [CI, Ϫ5.7 kg to 2.7 kg]; P
Ͼ 0.2). Never-
ment for baseline differences (Table 3
). This difference
theless, the difference in weight loss between the 2 diet
remained significant after weight loss amount was added to
groups for those who dropped out of the study was not
the model (P
ϭ 0.019), suggesting a direct effect of the
low-carbohydrate diet on glycemic control. However, thesignificance of the difference in the response of hemoglo-
bin A1c was not confirmed by an analysis that included
Changes in total and LDL cholesterol were not signif-
only the persons who completed the study (adjusted P
icantly different between groups (Table 3
0.080) or when baseline values were carried forward for
levels decreased more in the low-carbohydrate group than
missing persons (adjusted P
ϭ 0.18). Two persons on the
in the conventional diet group (P
ϭ 0.044 before and P
low-carbohydrate diet and 4 on the conventional diet de-
0.041 after adjustment for baseline variables) (Table 3
veloped diabetes at 1 year (P
separate sensitivity analysis that included only the 87 per-sons who completed the study confirmed the significance
Figure. Comparison of mean weight loss in kg between
of this finding (adjusted P
ϭ 0.016), as did the sensitivity
participants on the conventional diet and participants on the
analysis in which baseline values were carried forward for
low-carbohydrate diet at 6 months (n
؍ 118) and at 1 year
missing data (adjusted P
ϭ 0.001). Assignment to the low-
carbohydrate group (P
ϭ 0.003) and greater weight loss(P
ϭ 0.004) were each independent predictors of a de-crease in triglyceride concentration, suggesting a direct ef-fect of the low-carbohydrate diet on triglyceride reduction.
The HDL cholesterol concentration decreased more in
the conventional diet group than in the low-carbohydrate
group by 1 year (P
ϭ 0.025 before and P
ϭ 0.014 after
adjustment for baseline variables) (Table 3
). A separate
sensitivity analysis that included only the 87 persons who
completed the study confirmed the significance of this
finding (adjusted P
ϭ 0.004), as did the analysis using
baseline values carried forward for missing data (adjustedP
ϭ 0.011). The difference in mean HDL cholesterol re-
sponse between diet groups remained significant after ad-
ϭ 0.003 for comparisons between diet groups by random-coefficient
justment for both baseline variables and weight loss (P
analysis. The difference in weight loss was not significant between the 2diet groups by 1 year (P
ϭ 0.195 before and P
Ͼ 0.2 after adjustment
0.028), suggesting direct diet-related effects on HDL cho-
for baseline variables, by random-coefficient analysis). Error bars repre-
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10 783
ference in sodium intake between groups was statistically
As shown in Table 3
, changes in serum creatinine
significant and could represent less consumption of pre-
concentration did not significantly differ between groups.
packaged, low-fat but high-salt foods in the low-carbohy-
However, blood urea nitrogen level increased more in the
low-carbohydrate diet group. Changes in uric acid level
Despite modest and comparable overall weight loss,
the responses of triglycerides and HDL cholesterol to the
One person on the low-carbohydrate diet was hospi-
low-carbohydrate diet were more favorable than to the
talized with noncardiac chest pain during the third month
conventional diet. These findings are consistent with pre-
of the study. Two persons in the low-carbohydrate group
vious studies (7, 14 –17) and may be related to diminished
died, including one who died of complications of hyper-
very-low-density lipoprotein triglyceride production by the
osmolar coma 5 months into the study and another who
liver in response to decreased carbohydrate substrate deliv-
had severe ischemic cardiomyopathy and died suddenly 10
ery, as well as to improvements in insulin sensitivity. The
months after study enrollment. Laboratory values obtained
greater preservation of HDL cholesterol on the low-carbo-
14 days before this person’s death showed no electrolyte
hydrate diet may be a secondary effect of the greater de-
crease in triglycerides via cholesterol ester transferase orthrough downregulation of hepatic scavenger receptor B1levels (18). The expression of these receptors, which bind
HDL cholesterol and facilitate reverse cholesterol transport
To our knowledge, this is the largest and longest study
to the liver, may be modulated by dietary fats (18).
to date to compare weight and metabolic responses in per-
We also found that the hemoglobin A1c levels in the
sons with a high prevalence of diabetes or the metabolic
small group of persons with diabetes improved more on
syndrome receiving intensive counseling on either a low-
the low-carbohydrate diet, although this difference was not
carbohydrate diet or a conventional diet. We found no
significant in our sensitivity analyses. Nevertheless, the
significant difference in overall weight loss between persons
amount of improvement in hemoglobin A1c in the low-
on these 2 diets. In contrast to the findings by Foster and
carbohydrate group has a clinically significant effect on
colleagues (7), persons on the low-carbohydrate diet main-
micro- and macrovascular complications of diabetes (19).
tained most of their initial weight loss, whereas those on
Our study has several important limitations. Overall
the conventional diet continued to lose weight. Our differ-
weight loss was modest and the overall dropout rate was
ent findings may be due to the more intensive diet coun-
high. We tried to minimize any biasing effect by extracting
seling used in our study. We cannot exclude that a larger
1-year weights for persons who dropped out. Although
study might have demonstrated a statistically significant
these weights were not measured in a standardized fashion,
difference in weight loss between diets. Our enrollment
any random measurement errors would bias results toward
targets were based on a maximum anticipated weight loss
the null. Second, most persons did not meet their dietary
by 6 months and assumed weight stabilization thereafter.
targets (Ͻ30 g of carbohydrate daily in the low-carbohy-
Given that weight loss continued beyond 6 months in the
drate group and reduction of 500 calories per day in the
conventional diet group, we would have needed a sample
conventional diet). Meeting these targets would probably
size of approximately 284 persons per group to show a
have yielded different results. Last, there were some differ-
difference between groups at 1 year, assuming preservation
ences between the persons who completed the study and
those who dropped out, such as greater weight loss in the
Although it has been speculated that a low-carbohy-
former. Nevertheless, the observed differences in responses
drate diet would facilitate weight loss by promoting the
of triglyceride, HDL cholesterol, and hemoglobin A1c lev-
metabolism of adipose tissue (13), our data suggest that
els between diets were independent of differences in weight
weight loss differences may be explained by lower caloric
intake on a low-carbohydrate diet. If true, this may be
In summary, we found similar weight loss in persons
attributable to the simplicity of a low-carbohydrate diet or
randomly assigned to a low-carbohydrate diet or a conven-
to greater effects on satiety. Of note, persons on the low-
tional diet by 1 year. Despite modest overall weight loss in
carbohydrate diet who dropped out of the study were less
both diet groups, assignment to the low-carbohydrate
likely to lose weight, whereas those assigned to the conven-
group had a direct and more favorable effect on triglyceride
tional diet lost a similar amount of weight whether or not
level, HDL cholesterol level, and glycemic control in the
they remained in the study. This observation, together
smaller subgroup of patients with diabetes. These findings
with the difference between diets in weight loss beyond 6
give further evidence that restriction of carbohydrates in
months, raises the possibility that a low-carbohydrate diet
obese persons, who may be overconsuming carbohydrates
is less sustainable than a conventional diet. The low-carbo-
at baseline, may have favorable metabolic effects. Caution
hydrate diet followed in our study had healthy (lower so-
is still needed, however, in recommending a low-carbohy-
dium intake) and unhealthy (higher nonsaturated fat and
drate diet, as important concerns remain. Most important,
cholesterol levels and lower fiber intake) aspects. The dif-
future studies will need to evaluate whether a low-carbo-
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10
hydrate diet has more favorable effects on the development
A low-carbohydrate as compared with a low-fat diet in severe obesity. N Engl
of diabetes and on cardiovascular outcomes.
J Med. 2003;348:2074-81. [PMID: 12761364]
6. Bonow RO, Eckel RH.
Diet, obesity, and cardiovascular risk. N Engl J Med.
2003;348:2057-8. [PMID: 12761363]
From Philadelphia Veterans Affairs Medical Center, University of Penn-
7. Foster GD, Wyatt HR, Hill JO, McGuckin BG, Brill C, Mohammed BS, et
sylvania Medical Center, and Drexel University College of Medicine,
A randomized trial of a low-carbohydrate diet for obesity. N Engl J Med.
2003;348:2082-90. [PMID: 12761365]
8. Heshka S, Anderson JW, Atkinson RL, Greenway FL, Hill JO, Phinney SD,
The authors thank Dr. Stephen E. Kimmel and Dr.
Weight loss with self-help compared with a structured commercial program:
David Asch for their detailed review and comments on this manuscript,
a randomized trial. JAMA. 2003;289:1792-8. [PMID: 12684357]
as well as Dr. Justine Shults for her valuable consultation on the statis-
9. Friedewald WT, Levy RI, Fredrickson DS.
Estimation of the concentration of
low-density lipoprotein cholesterol in plasma, without use of the preparative ul-tracentrifuge. Clin Chem. 1972;18:499-502. [PMID: 4337382]
By the Veterans Affairs Healthcare Network Competi-
10. Executive Summary of The Third Report of The National Cholesterol Edu-
cation Program (NCEP) Expert Panel on Detection, Evaluation, And Treatmentof High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA. 2001;
Potential Financial Conflicts of Interest:
285:2486-97. [PMID: 11368702]
11. Twisk JW.
Applied Longitudinal Data Analysis for Epidemiology. A Practical
Requests for Single Reprints:
Frederick F. Samaha, MD, Cardiology,
Guide. Cambridge: Cambridge Univ Pr; 2003.
8th Floor, MC 111C, Philadelphia Veterans Affairs Medical Center,
12. Dawson-Saunders B, Trapp RG.
Basic and Clinical Biostatistics. Stamford,
University and Woodland Avenue, Philadelphia, PA 19104; e-mail,
13. Atkins RC.
Dr. Atkins’ New Diet Revolution. New York: Avon Books;
Current author addresses and author contributions are available at www
14. Lewis SB, Wallin JD, Kane JP, Gerich JE.
Effect of diet composition on
metabolic adaptations to hypocaloric nutrition: comparison of high carbohydrate
and high fat isocaloric diets. Am J Clin Nutr. 1977;30:160-70. [PMID: 835502]
15. Garg A, Grundy SM, Unger RH.
Comparison of effects of high and low
carbohydrate diets on plasma lipoproteins and insulin sensitivity in patients with
mild NIDDM. Diabetes. 1992;41:1278-85. [PMID: 1397701]
1. Mokdad AH, Serdula MK, Dietz WH, Bowman BA, Marks JS, Koplan JP.
16. Golay A, Eigenheer C, Morel Y, Kujawski P, Lehmann T, de Tonnac N.
The spread of the obesity epidemic in the United States, 1991-1998. JAMA.
Weight-loss with low or high carbohydrate diet? Int J Obes Relat Metab Disord.
2. Ford ES, Giles WH, Dietz WH.
Prevalence of the metabolic syndrome
17. Brehm BJ, Seeley RJ, Daniels SR, D’Alessio DA.
A randomized trial com-
among US adults: findings from the third National Health and Nutrition Exam-
paring a very low carbohydrate diet and a calorie-restricted low fat diet on body
ination Survey. JAMA. 2002;287:356-9. [PMID: 11790215]
weight and cardiovascular risk factors in healthy women. J Clin Endocrinol
3. National Heart, Lung, and Blood Institute, National Institute of Diabetes and
Digestive and Kidney Diseases. Obesity education initiative. In: Clinical Guide-
18. Hatahet W, Cole L, Kudchodkar BJ, Fungwe TV.
Dietary fats differentially
lines on the Identification, Education, and Treatment of Overweight and Obesity
modulate the expression of lecithin:cholesterol acyltransferase, apoprotein-A1 and
in Adults: The Evidence Report. NIH publication no. 98-4083. Bethesda, MD:
scavenger receptor b1 in rats. J Nutr. 2003;133:689-94. [PMID: 12612138]
National Heart, Lung, and Blood Institute in cooperation with the National
19. Stratton IM, Adler AI, Neil HA, Matthews DR, Manley SE, Cull CA, et al.
Institute of Diabetes and Digestive and Kidney Diseases; 1998.
Association of glycaemia with macrovascular and microvascular complications of
4. The truth about dieting. Consumer Reports. 2002;June:26-32.
type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321:
5. Samaha FF, Iqbal N, Seshadri P, Chicano KL, Daily DA, McGrory J, et al.
18 May 2004 Annals of Internal Medicine Volume 140 • Number 10 785
Current Author Addresses:
Drs. Stern, Iqbal, Seshadri, and Samaha,
Critical revision of the article for important intellectual content: L. Stern,
Ms. Chicano, Ms. Daily, Ms. McGrory, and Ms. Williams: Philadelphia
N. Iqbal, P. Seshadri, K.L. Chicano, D.A. Daily, J. McGrory, M. Wil-
Veterans Affairs Medical Center, Cardiology, 8th Floor MC 111C, Uni-
versity and Woodland Avenues, Philadelphia, PA 19104.
Final approval of the article: L. Stern, N. Iqbal, P. Seshadri, P. Seshadri,
Dr. Gracely: Family, Community, and Preventive Medicine, Drexel
K.L. Chicano, D.A. Daily, J. McGrory, M. Williams, E.J. Gracely, F.F.
University College of Medicine, 2900 Queen Lane, Philadelphia, PA
Provision of study materials or patients: L. Stern, N. Iqbal, P. Seshadri,K.L. Chicano, D.A. Daily, J. McGrory, M. Williams, F.F. Samaha.
Conception and design: L. Stern, N. Iqbal, P.
Statistical expertise: E.J. Gracely, F.F. Samaha.
Seshadri, K.L. Chicano, D.A. Daily, J. McGrory, M. Williams, E.J.
Obtaining of funding: L. Stern, N. Iqbal, P. Seshadri, K.L. Chicano,
D.A. Daily, J. McGrory, M. Williams, F.F. Samaha.
Analysis and interpretation of the data: P. Seshadri, E.J. Gracely, F.F.
Administrative, technical, or logistic support: F.F. Samaha.
Collection and assembly of data: L. Stern, N. Iqbal, P. Seshadri, K.L.
Drafting of the article: L. Stern, N. Iqbal, P. Seshadri, K.L. Chicano,
Chicano, D.A. Daily, J. McGrory, M. Williams, F.F. Samaha.
D.A. Daily, J. McGrory, M. Williams, F.F. Samaha.
Annals of Internal Medicine Volume • Number
LE ORIGINI DELLA LETTERATURA La nascita del e lingue romanze Le lingue volgari hanno origine in un periodo, durato oltre cinque secoli, che va dal crol o del ’Impero romano (476 DC) al 'anno 1000 circa. In quel periodo il latino, parlato da tutte le genti romanizzate, si trasforma di continuo poiché entra in contatto con i dialetti germanici (Franchi, Goti, Longobardi), il greco
COMPTE RENDU ANATOMOCYTOPATHOLOGIQUE 2 biopsies de muqueuse buccale : 2 sections pour chacune Néoformation tumorale de nature épithéliale en continuité multifocale avec l’épithélium malpighien muqueux, densément cellulaire, mal délimitée, à croissance nettement infiltrante, non encapsulée, ulcérée, procédant de cellules polygonales jointives agencées en lobules et cordo