Dietary protein and risk of ischemic heart disease in middle-aged
men1–3
Sarah Rosner Preis, Meir J Stampfer, Donna Spiegelman, Walter C Willett, and Eric B Rimm
ABSTRACT
Background: Prospective studies in US women have suggested an
inverse relation between dietary protein and risk of ischemic heart
disease (IHD). However, no large-scale prospective studies have
been conducted in US men.
Objective: The objective was to examine the association between
dietary protein and risk of IHD in a prospective study of US men.
Design: Intakes of protein and other nutrients were assessed by
using a validated food-frequency questionnaire at 4 time points
during follow-up of 43,960 men participating in the Health Profes-
sionals Follow-Up Study. Cox proportional hazards models were
used to calculate multivariable-adjusted relative risks (RRs) and
95% CIs.
Results: During 18 y of follow-up, we documented 2959 incident
cases of IHD. The RR of IHD was 1.08 (95% CI: 0.95, 1.23; P for
trend = 0.30) comparing the top with the bottom quintile of per-
centage of energy from total protein. RRs for animal and vegetable
protein were 1.11 (95% CI: 0.97, 1.28; P for trend = 0.18) and 0.93
(95% CI: 0.78, 1.12; P for trend = 0.49), respectively. When the
population was restricted to “healthy” men (those free of hyperten-
sion, hypercholesterolemia, and diabetes at baseline), the RR of IHD
was 1.21 (95% CI: 1.01, 1.44; P for trend = 0.02) for total protein, 1.25
(95% CI: 1.04, 1.51; P for trend = 0.02) for animal protein, and 0.93
(95% CI: 0.72, 1.19; P for trend = 0.65) for vegetable protein.
Conclusions: We observed no association between dietary protein and
risk of total IHD in this group of men aged 40–75 y. However, higher
intake of animal protein may be associated with an increased risk
of IHD in “healthy” men. Am J Clin Nutr 2010;92:1265–72.
INTRODUCTION
Several studies have suggested that a higher dietary intake of
protein in conjunction with a low- or moderate-carbohydrate diet
may be associated with lower blood pressure and cholesterol
concentrations and with short-term weight loss (1–3). However,
the relation between dietary protein and long-term risk of is-
chemic heart disease (IHD) has not been thoroughly studied.
With the current emphasis on higher-protein diets and weight
loss, it is important to determine the long-term ramifications of
such diets on chronic disease risk. Although previous results from
studies in US women suggest that higher-protein diets are as-
sociated with a lower risk of IHD (4, 5), similar results have not
been confirmed in US men. We examined the association between
dietary protein and risk of IHD in men participating in the Health
Professionals Follow-Up Study.
Am J Clin Nutr 2010;92:1265–72. Printed in USA. Ó 2010 American Society for Nutrition
SUBJECTS AND METHODS
Study population
Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015
The Health Professionals Follow-Up Study is an ongoing
prospective cohort study of 51,529 men aged 40–75 y at baseline
in 1986. The cohort participants are sent a biennial questionnaire
regarding diseases and lifestyle characteristics, such as smoking
status, medication use, and physical activity. Every 4 y, the
participants are sent a food-frequency questionnaire (FFQ) to
assess their diet composition. Approximately 94% of the cohort
has completed at least one follow-up questionnaire. We excluded
those who reported a history of myocardial infarction (MI),
angina, coronary artery bypass graft, other heart conditions,
stroke, pulmonary embolism, or cancer on the baseline ques-
tionnaire. In addition, those who had an implausible caloric
intake [,800 or .4200 kcal/d; cutoffs for extreme intake values
(64 SD) derived from the baseline FFQ] or had .70 missing
responses to food items were excluded, which resulted in
a baseline population of 43,960 for the current analysis. This
study was approved by the Harvard Institutional Review Board.
Dietary assessment
Diet was assessed at baseline in 1986 and in 1990, 1994, 1998,
and 2002 by using a 131-item FFQ. Details of the method of
calculating nutrient intakes from the FFQ were previously de-
scribed (6). Intakes of total, animal, and vegetable protein were
calculated for each participant. Protein intake was expressed as
a percentage of energy by multiplying the grams of protein
consumed per day by the number of kilocalories in 1 g of protein
(4 kcal/g) and then dividing by the subject’s total caloric intake
(7). The other macronutrients, carbohydrate, and fats (saturated,
monounsaturated, polyunsaturated, and trans) were also ex-
1
From the Departments of Epidemiology (SRP, MJS, DS, WCW, and
EBR), Nutrition (MJS, WCW, and EBR), and Biostatistics (DS), Harvard
School of Public Health, Boston, MA, and the Channing Laboratory, De-
partment of Medicine, Brigham and Women’s Hospital and Harvard Medical
School, Boston, MA (MJS, DS, WCW, and EBR).
2
Supported by grants from the National Institutes of Health (HL35464
and CA55075) and by the Kirschstein-NRSA Aging Training Grant
(AG000158).
3
Address correspondence to EB Rimm, Harvard School of Public Health,
655 Huntington Avenue, Building II, Room 373a, Boston, MA 02115.
E-mail: erimm@hsph.harvard.edu.
Received May 6, 2010. Accepted for publication August 25, 2010.
First published online September 29, 2010; doi: 10.3945/ajcn.2010.29626.
1265
1266 PREIS ET AL
pressed as a percentage of energy, assuming 4 kcal per gram of
carbohydrate and 9 kcal per gram of fat. In addition to the
macronutrients, nutrient intakes were calculated for glycemic
index, fiber, folate, vitamin C, magnesium, total omega-3 (n23)
fatty acids, and alcohol. With the exception of alcohol, all nu-
trients were energy-adjusted by using the residual method (8).
The validity of the 131-item FFQ was assessed in a subsample
of the baseline study population (9). The validity of the FFQ was
assessed by comparing the nutrient intakes estimated by the
questionnaire with the intakes calculated from an average of the 2
weighed 1-wk diet records. The deattenuated energy-adjusted
Pearson correlation coefficient for the macronutrients between
the diet records and the FFQ was 0.67 for fat, 0.73 for carbo-
hydrate, and 0.44 for protein.
Assessment of IHD endpoints
The primary endpoints of interest were incident nonfatal MI
and fatal IHD occurring between the return of the 1986 ques-
tionnaire and 31 January 2004. Nonfatal MI was assessed bi-
ennially with a mailed questionnaire. If a participant reported
a diagnosis and hospitalization for MI, we first obtained con-
firmation of the event and consent by letter or by phone to review
their medical records. A physician blinded to exposure status
verified the report of MI through a review of medical or hospital
records using the World Health Organization criteria of symp-
toms and either typical electrocardiographic changes or elevated
cardiac enzymes (10).
Deaths were ascertained by contact with family members or
through the National Death Index. Fatal IHD was confirmed from
the medical records or autopsy reports or if IHD was listed as the
cause of death on the death certificate and there was evidence of
previous IHD in the records.
Statistical analysis
A multivariate nutrient density model was used to analyze the
association between protein intake and risk of IHD (8, 11). Our
purpose in using the percentage of energy from protein instead of
grams of protein was to examine the effect of increasing protein
intake independent of increases in total energy intake. The co-
efficient for protein reflects the substitution of an equal amount of
energy from protein for carbohydrate rather than an absolute
increase in protein intake. The age-adjusted nutrient density
model contains quintiles of percentage of energy from protein,
percentage of energy from saturated fat, percentage of energy
from monounsaturated fat, percentage of energy from poly-
unsaturated fat, percentage of energy from trans fat, and total
energy intake. The multivariate nutrient density model contains
all variables in the age-adjusted model plus variables for other
established risk factors for IHD including body mass index
(BMI, in kg/m2; ,23, 23–24.9, 25–28.9, or 29), cigarette
smoking (never smoker, nonsmoker with unknown past history,
past smoker, or current smoker of 1–14, 15–24, 25, or an
unknown number of cigarettes daily), parental history of MI
before age 65 y (yes or no), alcohol consumption (0, 0.1–4.9, 5–
14.9, or 15 g/d), multivitamin use (yes or no), and quintiles of
physical activity (metabolic equivalents/d), glycemic index, fo-
late (lg/d), fiber (g/d), vitamin C (mg/d), magnesium (mg/d),
and total omega-3 fatty acids (g/d) (12, 13). Because protein
intake was shown to be inversely associated with IHD mortality
in a previous study, we repeated our main analysis examining
nonfatal and fatal IHD endpoints separately (5).
Because the participants who developed hypertension, di-
abetes, or hypercholesterolemia before the start of the study may
have altered their diet after their diagnosis, we fit an alternative
multivariate nutrient-density model that contained baseline
(1986) status of hypertension, diabetes, and hypercholesterol-
emia in addition to all of the abovementioned variables. We did
not control for the development of hypertension, diabetes, and
hypercholesterolemia during the study because these variables
are potential intermediates in the causal pathway between dietary
protein intake and IHD (14).
Each participant contributed person-time to the analysis,
starting from the date of the return of their 1986 questionnaire
until 31 January 2004, death, loss to follow-up, diagnosis of
cancer or stroke, or development of IHD, whichever occurred
Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015
first. Incidence rates for MI were calculated for each quintile of
percentage of energy from protein. A Cox proportional hazards
regression model, stratified jointly by age in months and by each
of the eight 2-y follow-up time periods, was fit by using PROC
PHREG in SAS (version 9.1; SAS Institute Inc, Cary, NC) to
calculate incidence rate ratios (relative risks; RRs) and 95% CIs.
To assess the presence of a linear trend, a median score variable
was constructed by using the median protein intake of each
quintile and its significance was assessed by using a Wald test. All
P values are 2-sided.
For the repeated measurements of dietary protein, we used the
cumulative average approach to assign an individual’s intake at
each time period, which gives greater weight to more recent diet
(11). The cumulative average approach minimizes measurement
error because it uses all previous dietary assessments during
follow-up (11). If a person developed an intermediate event that
could alter their diet (eg, hypercholesterolemia, hypertension,
angina, diabetes, and cancer), only their diet before the de-
velopment of the event was considered in the analysis.
The interaction of total protein intake with traditional car-
diovascular disease (CVD) risk factors was examined by in-
cluding a cross-product term for percentage of energy from
protein multiplied by the risk factor, and its significance was
assessed by using a likelihood ratio test. The risk factors con-
sidered were hypertension (yes or no), hypercholesterolemia (yes
or no), diabetes (yes or no), BMI (in kg/m2; 25 or ,25), and
glycemic index (55 or ,55).
RESULTS
Baseline characteristics
During the 18 y (688,455 person-years) of follow-up, we
documented 2959 incident IHD events (1804 nonfatal MIs and
1155 fatal cases of IHD) among the 43,960 participants included
in the analysis. Baseline characteristics of the study population
according to quintile of percentage of energy from total protein
are presented in Table 1. Men with a higher percentage of en-
ergy from protein were more likely to have reported a diagnosis
of hypertension, hypercholesterolemia, and diabetes and also
had higher average folate, vitamin C, magnesium, and omega-3
fatty acid intakes (P for linear trend ,0.0001 for all variables).
Those in the top quintile also had lower alcohol consumption
and lower total energy intake (P for linear trend ,0.0001 for
both variables).
 DIETARY PROTEIN AND RISK OF ISCHEMIC HEART DISEASE 1267
TABLE 1
Distribution of ischemic heart disease risk factors according to quintile (Q) of percentage of energy from total protein at baseline (1986): Health
Professionals Follow-Up Study (n = 43,960), 1986–20041
Percentage of energy from total protein

Q1 Q3 Q5 P for linear trend2

Age (y) 53 6 103 53 6 9 55 6 9 ,0.0001

Smoking status [n (%)]
Never 3993 (45.2) 4007 (47.5) 4409 (44.9) Referent
Past 3405 (38.6) 3336 (39.5) 4126 (42.1) 0.009
1–14 cigarettes/d 262 (3.0) 216 (2.6) 251 (2.6) 0.07
15–24 cigarettes/d 334 (3.8) 294 (3.5) 276 (2.8) ,0.0001
25 cigarettes/d 390 (4.4) 214 (2.5) 209 (2.1) ,0.0001
Unknown no. of cigarettes/d 105 (1.2) 72 (0.9) 96 (1.0) 0.14
Missing 337 (3.8) 301 (3.6) 446 (4.5) 0.03
History of hypertension [n (%)] 1822 (20.6) 1812 (21.5) 2512 (25.6) ,0.0001
History of hypercholesterolemia [n (%)] 807 (9.1) 873 (10.3) 1227 (12.5) ,0.0001

Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015

History of diabetes [n (%)] 114 (1.3) 147 (1.7) 506 (5.2) ,0.0001
Parental history of MI ,65 y [n (%)] 1025 (11.6) 975 (11.6) 1283 (13.1) 0.0004
Exercise (METs) 20.9 6 30.1 20.8 6 27.4 21.1 6 29.1 0.89
BMI (kg/m2) 25.1 6 3.2 25.5 6 3.3 25.9 6 3.6 ,0.0001
Calories (kcal/d) 2137 6 664 2024 6 597 1780 6 574 ,0.0001
Total protein (% of energy) 14.2 6 1.4 18.2 6 0.4 23.2 6 2.3 ,0.0001
Animal protein (% of energy) 9.3 6 1.8 13.2 6 1.2 18.4 6 2.7 ,0.0001
Vegetable protein (% of energy) 4.9 6 1.3 5.0 6 1.2 4.9 6 1.3 0.60
Carbohydrates (% of energy) 50.9 6 9.3 46.9 6 7.5 42.9 6 7.9 ,0.0001
Saturated fat (% of energy) 10.5 6 2.8 11.2 6 2.7 11.1 6 2.9 ,0.0001
Monounsaturated fat (% of energy) 11.8 6 2.8 12.5 6 2.6 12.2 6 2.8 ,0.0001
Polyunsaturated fat (% of energy) 5.9 6 1.8 6.0 6 1.5 6.0 6 1.5 ,0.0001
trans Fat (% of energy) 1.4 6 0.6 1.3 6 0.5 1.1 6 0.4 ,0.0001
Alcohol (% of energy) 5.9 6 7.4 3.9 6 4.8 2.7 6 3.6 ,0.0001
Alcohol (g/d) 17.5 6 22.0 10.9 6 13.5 6.8 6 9.1 ,0.0001
Folate, energy-adjusted (lg/d) 442 6 255 476 6 262 520 6 312 ,0.0001
Fiber, energy-adjusted (g/d) 19.9 6 7.6 21.0 6 6.6 21.6 6 7.5 ,0.0001
Vitamin C, energy-adjusted (mg/d) 389 6 423 421 6 469 488 6 532 ,0.0001
Magnesium, energy-adjusted (mg/d) 324 6 81 351 6 75 382 6 92 ,0.0001
Omega-3 fatty acids, energy-adjusted (g/d) 1.3 6 0.4 1.4 6 0.4 1.6 6 0.5 ,0.0001
Glycemic index 54.0 6 3.8 53.3 6 3.3 52.0 6 3.9 ,0.0001
1
METs, metabolic equivalent tasks; MI, myocardial infarction.
2
P value for linear trend was calculated across all 5 quintiles by using logistic regression for dichotomous variables and linear regression for continuous
variables.
3
Mean 6 SD (all such values).

Substitution of protein for carbohydrate
The association between quintile of percentage of energy from
protein and risk of IHD is shown in Table 2. For total protein, the
RR for the comparison of those in the top total protein quintile
(median: 22.5% of energy) with those in the bottom quintile
(median: 14.6% of energy) was 1.22 (95% CI: 1.08, 1.38) after
adjustment for age, dietary fat, and total energy intake. This
model substituted protein for an isocaloric amount of carbohy-
drate. Additional adjustment for BMI; quintiles of fiber, folate,
vitamin C, magnesium, and omega-3 fatty acids; glycemic in-
dex; physical activity; family history of MI; cigarette smoking;
and alcohol and multivitamin use resulted in an RR of 1.20
(95% CI: 1.05, 1.37; P for trend = 0.006). After additional ad-
justment for baseline status of hypertension, hypercholesterol-
emia, and diabetes, the RR decreased to 1.08 (95% CI: 0.95,
1.23), and the P for trend was 0.30.
The Cox proportional hazards models for animal protein
showed results similar to those for total protein. The age- and
macronutrient-adjusted RR for the comparison of those in the top
quintile of animal protein (median: 17.7% of energy) intake with
those in the bottom (median: 9.3% of energy) was 1.27 (95% CI:
1.12, 1.44). In the fully adjusted model, the RR for the top
quintile was 1.11 (95% CI: 0.97, 1.28), and the P for trend was
0.18. For vegetable protein, the age- and macronutrient-adjusted
RR for the comparison of those in the top quintile (median:
6.5% of energy) with those in the bottom quintile (median: 3.7%
of energy) was 0.86 (95% CI: 0.75, 0.98). In the fully adjusted
model, the RR for the top quintile was 0.93 (95% CI: 0.78,
1.12), and the P for trend was 0.49.
Exclusion of participants with baseline conditions
Some participants had hypertension, diabetes, and hyper-
cholesterolemia at baseline, which may have led to a change in
their diet before the onset of the study. Control for these potential
confounders attenuated the main results (Table 2). However, it is
unclear whether the attenuation was due to the removal of
1268 PREIS ET AL
TABLE 2
Relative risks (RRs) and 95% CIs for total ischemic heart disease (IHD) according to quintile (Q) of percentage of energy from protein: Health Professionals
Follow-Up Study (n = 43,960), 1986–20041
Percentage of energy from protein

Q1 Q2 Q3 Q4 Q5 P for trend

Total protein
Median (% of energy) 14.6 16.7 18.2 19.8 22.5
Total IHD (no. of cases) 552 548 572 584 703
Multivariate RR
Model 1 1.00 (referent) 0.99 (0.87, 1.11) 1.03 (0.91, 1.16) 1.03 (0.91, 1.16) 1.22 (1.08, 1.38) 0.0005
Model 2 1.00 (referent) 1.03 (0.91, 1.17) 1.09 (0.96, 1.23) 1.08 (0.95, 1.22) 1.20 (1.05, 1.37) 0.006
Model 3 1.00 (referent) 1.03 (0.91, 1.16) 1.07 (0.94, 1.21) 1.03 (0.90, 1.16) 1.08 (0.95, 1.23) 0.30
Animal protein
Median (% of energy) 9.3 11.5 13.1 14.9 17.7
Total IHD (no. of cases) 527 547 549 600 736
Multivariate RR2

Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015

Model 1 1.00 (referent) 1.04 (0.92, 1.18) 1.01 (0.89, 1.15) 1.07 (0.95, 1.21) 1.27 (1.12, 1.44) 0.0001
Model 2 1.00 (referent) 1.06 (0.94, 1.20) 1.05 (0.92, 1.19) 1.09 (0.96, 1.24) 1.24 (1.08, 1.42) 0.003
Model 3 1.00 (referent) 1.05 (0.93, 1.16) 1.02 (0.90, 1.16) 1.03 (0.90, 1.17) 1.11 (0.97, 1.28) 0.18
Vegetable protein
Median (% of energy) 3.7 4.4 4.9 5.5 6.5
Total IHD (no. of cases) 704 600 570 564 521
Multivariate RR3
Model 1 1.00 (referent) 0.90 (0.80, 1.01) 0.87 (0.77, 0.98) 0.86 (0.76, 0.97) 0.86 (0.75, 0.98) 0.02
Model 2 1.00 (referent) 0.96 (0.85, 1.08) 0.94 (0.82, 1.08) 0.95 (0.81, 1.11) 0.96 (0.80, 1.15) 0.72
Model 3 1.00 (referent) 0.96 (0.85, 1.08) 0.94 (0.82, 1.08) 0.94 (0.80, 1.09) 0.93 (0.78, 1.12) 0.49
1
A Cox proportional hazards model was used to calculate RRs and 95% CIs, and P values were derived from a Wald test. Model 1 was adjusted for age
and quintiles of percentage of energy from saturated fat, monounsaturated fat, polyunsaturated fat, trans fat, and calories. Model 2 was adjusted as for model 1
plus quintiles of fiber, folate, vitamin C, magnesium, total omega-3 fatty acids, glycemic index, physical activity, family history of myocardial infarction (yes
or no), BMI (in kg/m2; ,23, 23–24.9, 25–28.9, or 29), cigarette smoking (never, nonsmoker with unknown past history, past smoker, or current smoker of 1–
14, 15–24, 25, or an unknown number of cigarettes daily), alcohol use (0, 0.1–4.9, 5–14.9, or 15 g/d), and multivitamin use (yes or no). Model 3 was
adjusted as for model 2 plus baseline (1986) status of hypertension, hypercholesterolemia, and diabetes.
2
Additionally adjusted for quintiles of percentage of energy from vegetable protein.
3
Additionally adjusted for quintiles of percentage of energy from animal protein.

confounding or to the control of potential intermediate variables
in the causal pathway between protein intake and IHD occur-
rence. We also repeated the analysis excluding those who
reported a diagnosis of hypertension, diabetes, or hypercholes-
terolemia at baseline and found that the results were further from
the null than those for the full cohort. For example, the RR
between extreme quintiles was 1.21 (95% CI: 1.01, 1.44; P for
trend = 0.02) for total protein, 1.25 (95% CI: 1.04, 1.51; P for
trend = 0.02) for animal protein, and 0.93 (95% CI: 0.72, 1.19; P
for trend = 0.65) for vegetable protein. These results are similar
to those obtained with the multivariate models of the main
analysis, which did not control for baseline status of hyperten-
sion, diabetes, and hypercholesterolemia.
Dietary protein and fatal and nonfatal IHD
We conducted further subanalyses to examine the association
between dietary protein intake and risk of nonfatal MI and fatal
IHD separately (Table 3). For both nonfatal MI and fatal IHD,
the results for total protein and animal protein were similar to
those for total IHD. For vegetable protein, the RR for the
comparison of those in the top quintile with those in the bottom
quintile was 1.18 (95% CI: 0.93, 1.48) in the fully adjusted
model for nonfatal MI and 0.66 (95% CI: 0.49, 0.88) with a P
for trend of 0.005 for fatal IHD.
Effect modification by CVD risk factors
No statistically significant interaction between diabetes, hy-
percholesterolemia, or high BMI and total, animal, or vegetable
protein on the risk of IHD was observed. However, we did find
significant differences between protein and risk of IHD by
baseline hypertension and average glycemic index of the diet.
The increased risk of IHD associated with higher total and animal
protein was strongest among men without hypertension (Table
4). Total and animal protein were more strongly associated with
risk of MI among men consuming a diet with a lower glycemic
index (,55) (Table 4).
DISCUSSION
In this study we examined the risk of IHD associated with the
substitution of an equal percentage of energy from total, animal,
and vegetable protein for carbohydrate in a large prospective
cohort of US men. We found no association between quintiles of
percentage of energy from total, animal, or vegetable protein and
risk of IHD across the range of intakes. We found a significant
inverse association between higher intake of vegetable protein
and risk of fatal IHD. In addition, when the study population was
restricted to men free of diabetes, hypertension, and hypercho-
lesterolemia, a higher intake of total and animal protein was
associated with an increased risk of IHD. Finally, total and animal
 DIETARY PROTEIN AND RISK OF ISCHEMIC HEART DISEASE 1269
TABLE 3
Relative risks (RRs) and 95% CIs for nonfatal myocardial infarction (MI) and fatal ischemic heart disease (IHD) according to quintile (Q) of percentage of
energy from protein: Health Professionals Follow-Up Study (n = 43,960), 1986–20041
Percentage of energy from protein

Q1 Q2 Q3 Q4 Q5 P for trend

Total protein
Median (% of energy) 14.6 16.7 18.2 19.8 22.5
Nonfatal MI (no. of cases) 341 337 361 356 409
Multivariate RR
Model 1 1.00 (referent) 0.97 (0.84, 1.14) 1.03 (0.89, 1.20) 1.02 (0.87, 1.18) 1.17 (1.00, 1.36) 0.03
Model 2 1.00 (referent) 1.01 (0.87, 1.18) 1.07 (0.91, 1.25) 1.06 (0.90, 1.24) 1.19 (1.00, 1.41) 0.04
Model 3 1.00 (referent) 1.01 (0.86, 1.18) 1.05 (0.90, 1.23) 1.02 (0.87, 1.20) 1.10 (0.92, 1.30) 0.30
Fatal IHD (no. of cases) 211 211 211 228 294
Multivariate RR
Model 1 1.00 (referent) 1.01 (0.83, 1.22) 1.03 (0.84, 1.25) 1.06 (0.87, 1.28) 1.30 (1.08, 1.57) 0.004
Model 2 1.00 (referent) 1.05 (0.87, 1.29) 1.12 (0.92, 1.38) 1.10 (0.90, 1.35) 1.22 (0.98, 1.50) 0.07

Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015

Model 3 1.00 (referent) 1.05 (0.86, 1.28) 1.08 (0.88, 1.33) 1.02 (0.83, 1.25) 1.05 (0.85, 1.30) 0.79
Animal protein
Median (% of energy) 9.3 11.5 13.1 14.9 17.7
Nonfatal MI (no. of cases) 336 336 340 376 416
Multivariate RR2
Model 1 1.00 (referent) 0.99 (0.85, 1.16) 0.98 (0.83, 1.14) 1.07 (0.91, 1.25) 1.19 (1.01, 1.39) 0.02
Model 2 1.00 (referent) 1.02 (0.87, 1.19) 1.01 (0.86, 1.18) 1.10 (0.93, 1.29) 1.21 (1.01, 1.44) 0.02
Model 3 1.00 (referent) 1.01 (0.86, 1.18) 0.99 (0.84, 1.16) 1.05 (0.89, 1.23) 1.12 (0.94, 1.33) 0.18
Fatal IHD (no. of cases) 191 211 209 224 320
Multivariate RR2
Model 1 1.00 (referent) 1.13 (0.92, 1.37) 1.08 (0.88, 1.32) 1.08 (0.88, 1.32) 1.40 (1.15, 1.71) 0.002
Model 2 1.00 (referent) 1.15 (0.93, 1.40) 1.12 (0.91, 1.38) 1.06 (0.86, 1.31) 1.28 (1.03, 1.59) 0.06
Model 3 1.00 (referent) 1.12 (0.91, 1.37) 1.08 (0.87, 1.33) 0.98 (0.79, 1.21) 1.10 (0.88, 1.37) 0.71
Vegetable protein
Median (% of energy) 3.7 4.4 4.9 5.5 6.5
Nonfatal MI (no. of cases) 392 370 345 353 344
Multivariate RR3
Model 1 1.00 (referent) 0.98 (0.85, 1.14) 0.96 (0.82, 1.11) 1.00 (0.85, 1.17) 1.07 (0.90, 1.28) 0.43
Model 2 1.00 (referent) 1.07 (0.91, 1.26) 1.04 (0.87, 1.25) 1.11 (0.91, 1.35) 1.21 (0.96, 1.53) 0.10
Model 3 1.00 (referent) 1.07 (0.91, 1.25) 1.03 (0.86, 1.24) 1.09 (0.89, 1.33) 1.18 (0.93, 1.48) 0.18
Fatal IHD (no. of cases) 312 230 225 211 177
Multivariate RR3
Model 1 1.00 (referent) 0.79 (0.66, 0.95) 0.76 (0.64, 0.92) 0.69 (0.57, 0.84) 0.60 (0.48, 0.75) ,0.0001
Model 2 1.00 (referent) 0.84 (0.69, 1.02) 0.83 (0.67, 1.03) 0.76 (0.60, 0.97) 0.67 (0.50, 0.90) 0.009
Model 3 1.00 (referent) 0.84 (0.69, 1.02) 0.84 (0.68, 1.05) 0.76 (0.59, 0.96) 0.66 (0.49, 0.88) 0.005
1
A Cox proportional hazards model was used to calculate RRs and 95% CIs, and P values were derived from a Wald test. Model 1 was adjusted for age
and quintiles of percentage of energy from saturated fat, monounsaturated fat, polyunsaturated fat, trans fat, and calories. Model 2 was adjusted as for model 1
plus quintiles of fiber, folate, vitamin C, magnesium, total omega-3 fatty acids, glycemic index, physical activity, family history of myocardial infarction (yes
or no), BMI (in kg/m2; ,23, 23–24.9, 25–28.9, or 29), cigarette smoking (never, nonsmoker with unknown past history, past smoker, or current smoker of 1–
14, 15–24, 25, or an unknown number of cigarettes daily), alcohol (0, 0.1–4.9, 5–14.9, or 15 g/d), and multivitamin use (yes or no). Model 3 was adjusted
as for model 2 plus baseline (1986) status of hypertension, hypercholesterolemia, and diabetes.
2
Additionally adjusted for quintiles of percentage of energy from vegetable protein.
3
Additionally adjusted for quintiles of percentage of energy from animal protein.

protein intakes were more strongly associated with risk of MI
among men without hypertension and among men consuming
a diet with a lower glycemic index (,55).
In the context of the current literature
The Kuopio Ischemic Heart Disease Risk Factor Study, which
was conducted in a Finnish cohort of ’2000 men, reported on
the association between dietary protein and risk of IHD (15).
The age- and examination year–adjusted RR for the comparison
of extreme quartiles of percentage of energy from total protein
was 0.78 (95% CI: 0.56, 1.09). Consistent with their age-
adjusted RR, our multivariate results also showed no statistically
significant association between total protein and IHD risk.
Multivariate RRs are not presented in their article.
The association between dietary protein and risk of IHD has
been investigated in 2 prior cohort studies in all-female US
populations. In 20 y of follow-up in the Nurses’ Health Study, the
authors found a RR of 1.06 (95% CI: 0.86, 1.30) for total IHD
when comparing extreme deciles of percentage energy from total
protein (16). The corresponding RR for animal protein was 1.13
(95% CI: 0.91, 1.41) and 1.08 (95% CI: 0.82, 1.43) for vegetable
protein. Our results are consistent with those from the Nurses’
Health Study cohort which had a similar range of protein intake
as our cohort.
1270 PREIS ET AL
TABLE 4
Relative risks (and 95% CIs) of ischemic heart disease according to quintile (Q) of percentage of energy from protein, stratified by hypertension and
glycemic index: Health Professionals Follow-Up Study, 1986–20041
Percentage of energy from protein

Q1 Q2 Q3 Q4 Q5 P for trend P for interaction

Total protein
Hypertension
No 1.00 (referent) 1.02 (0.85, 1.22) 1.13 (0.94, 1.36) 1.05 (0.86, 1.27) 1.25 (1.02, 1.53) 0.04 0.05
Yes 1.00 (referent) 1.06 (0.89, 1.26) 1.04 (0.87, 1.23) 1.01 (0.85, 1.20) 0.97 (0.81, 1.16) 0.57
Glycemic index2
Low (,55) 1.00 (referent) 1.10 (0.94, 1.29) 1.16 (1.00, 1.36) 1.13 (0.97, 1.32) 1.20 (1.02, 1.41) 0.04 0.03
High (55) 1.00 (referent) 0.90 (0.73, 1.11) 0.91 (0.73, 1.14) 0.85 (0.66, 1.08) 0.86 (0.66, 1.12) 0.21
Animal protein3
Hypertension
No 1.00 (referent) 1.06 (0.88, 1.27) 1.18 (0.96, 1.42) 1.05 (0.86, 1.28) 1.27 (1.03, 1.56) 0.05 0.04
Yes 1.00 (referent) 1.05 (0.89, 1.25) 0.88 (0.74, 1.06) 1.01 (0.85, 1.20) 0.98 (0.82, 1.18) 0.82

Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015

Glycemic index2
Low (,55) 1.00 (referent) 1.07 (0.92, 1.26) 1.07 (0.91, 1.26) 1.11 (0.95, 1.30) 1.19 (1.00, 1.40) 0.04 0.03
High (55) 1.00 (referent) 1.05 (0.85, 1.29) 0.93 (0.74, 1.17) 0.87 (0.68, 1.11) 0.99 (0.75, 1.30) 0.55
Vegetable protein4
Hypertension
No 1.00 (referent) 0.90 (0.74, 1.08) 0.91 (0.73, 1.12) 0.89 (0.70, 1.13) 0.89 (0.68, 1.18) 0.51 0.42
Yes 1.00 (referent) 0.97 (0.82, 1.15) 0.92 (0.77, 1.11) 0.93 (0.76, 1.15) 0.89 (0.70, 1.13) 0.33
Glycemic index2
Low (,55) 1.00 (referent) 1.05 (0.90, 1.21) 1.03 (0.87, 1.21) 0.99 (0.82, 1.19) 0.96 (0.77, 1.19) 0.57 0.74
High (55) 1.00 (referent) 0.77 (0.61, 0.98) 0.74 (0.56, 0.97) 0.78 (0.58, 1.04) 0.85 (0.60, 1.19) 0.49
1
A Cox proportional hazards model was used to calculate RRs and 95% CIs, and P values were derived from a Wald test. The number of cases is 1318
for nonhypertensive, 1638 for hypertensive, 2104 for low glycemic index, and 852 for high glycemic index. All models were adjusted for age; quintiles of
percentage of energy from saturated fat, monounsaturated fat, polyunsaturated fat, trans fat, calories, fiber, folate, vitamin C, magnesium, and total omega-3
fatty acids; glycemic index; physical activity; family history of myocardial infarction (yes or no); BMI (in kg/m2; ,23, 23–24.9, 25–28.9, or 29); cigarette
smoking (never, nonsmoker with unknown past history, past smoker, or current smoker of 1–14, 15–24, 25, or an unknown number of cigarettes daily);
alcohol use (0, 0.1–4.9, 5–14.9, or 15 g/d); multivitamin use (yes or no); and baseline status of hypercholesterolemia and diabetes.
2
Additionally adjusted for baseline hypertension status. Model was adjusted for glycemic index (continuous) instead of glycemic index in quintiles.
3
Additionally adjusted for quintiles of percentage of energy from vegetable protein.
4
Additionally adjusted for quintiles of percentage of energy from animal protein.

The association between dietary protein and risk of IHD
mortality was also examined in the Iowa Women’s Health Study
(5). In a comparison of extreme quintiles of protein intake, the RR
of fatal IHD was 0.84 (95% CI: 0.39, 1.79) for total protein
(22.0% compared with 14.1% of energy), 0.88 (95% CI: 0.42,
1.86) for animal protein (17.5% compared with 8.9% of energy),
and 0.70 (95% CI: 0.49, 0.99) for vegetable protein (6.1%
compared with 3.7% of energy). Our results for fatal IHD were
similar to the results found in the Iowa Women’s Health Study, in
which we found a significant inverse association for vegetable
protein and no association of fatal IHD with total or animal
protein.
Potential biological mechanisms
If dietary protein intake, especially vegetable protein intake,
reduces the risk of IHD, it may be mediated through beneficial
effects on blood pressure, cholesterol, and body weight. Ran-
domized clinical trials have shown an inverse association
between higher vegetable protein intake and blood pressure (1,
17–20). Observational studies have also shown an inverse relation
between nonanimal protein intake and blood pressure (21, 22). In
addition, the effect of protein may depend on the quality of
carbohydrates that is being replaced, because we observed
a differential effect of total and animal protein depending on the
average glycemic index of a participant’s diet.
With respect to dietary protein and plasma cholesterol, in
a meta-analysis of 38 clinical trials of soy-protein diets, there was
an average decrease of 9.3% for total cholesterol, 12.9% for LDL
cholesterol, and 10.5% for triglycerides and an increase of 2.4%
for HDL cholesterol when compared with subjects consuming
a control diet (high animal protein) (23). However, more recent
studies have suggested that the effect of soy on LDL cholesterol is
much smaller (24). Other randomized trials that compared high-
protein with control diets have shown more favorable cholesterol
concentrations for the participants assigned to the high-protein
diets (2, 19, 25–29).
A third method by which dietary protein can reduce the risk of
MI is through its effect on weight loss. Several randomized trials
have shown that individuals assigned to a low-carbohydrate,
high-protein diet had a greater weight loss than did those assigned
to a low-fat diet; however, few studies found statistically sig-
nificant differences between groups after long-term follow-up (2,
25, 26, 30–34). Foods high in vegetable protein, such as nuts and
legumes, have been shown to be inversely associated with risk of
IHD (35, 36). These sources of protein also have a high content of
unsaturated fatty acids, which have been shown to be especially
beneficial at preventing sudden cardiac death (37). However, we
controlled for both fats and omega-3 fatty acids in our analysis
and still found an inverse association between vegetable protein
and fatal IHD. In general, vegetable protein has a lower content
 DIETARY PROTEIN AND RISK OF ISCHEMIC HEART DISEASE
of essential amino acids—namely methionine, lysine, and
tryptophan—than does animal protein (38). Vegetable protein
contains a higher content of the nonessential amino acids argi-
nine, glycine, alanine, and serine. Intake of essential amino acids
results in increased insulin release to stimulate protein syn-
thesis and storage, whereas intake of nonessential amino acids
results in gluconeogenesis (38). However, a higher intake of the
amino acid arginine may increase concentrations of the vaso-
dilator nitric oxide and decrease blood pressure (1). Also, ar-
ginine has been shown to have a powerful depressor effect when
administered intravenously (39). One study suggested that higher
intakes of glutamic acid, the primary amino acid found in veg-
etable protein, were associated with lower blood pressure (40).
Thus, the stronger beneficial effects of vegetable protein on risk
of IHD may be explained through these mechanisms.
Although it is biologically plausible that a higher intake of
dietary protein, especially from vegetable sources, may decrease
the risk of IHD, it is unclear why we observed an inverse as-
sociation only for fatal MI and not for total or nonfatal IHD. Our
results are consistent with a previous study that showed a statis-
tically significant inverse association between vegetable protein
and IHD mortality; however, the biological mechanism to explain
this inverse association is unclear (5). It is important to note that we
were unable to control for several factors related to health con-
sciousness (eg, such as medication compliance and behavior
change after nonfatal disease), which may have confounded the
results for fatal IHD. It is also possible that this relation may have
arisen by chance; therefore, confirmation in other prospective
studies is men is necessary.
Limitations
We recognize that assessment of dietary protein through self-
administered FFQs has limitations. Misclassification could occur
if there were dietary changes that were not captured during the 4-y
period between the FFQ administrations. This misclassification is
likely to be nondifferential, causing the effect estimates to be
biased toward the null. In our dietary validation study, the cor-
relation between protein intake estimated from the FFQ and
protein intake estimated from the diet records was relatively low
(0.44), which suggested that the FFQ may not be an ideal measure
of protein intake (9). However, in the validation study, the range
of protein intakes was limited (10–25% of energy), which also
likely contributed to the modest correlation. The same FFQ was
used in previous studies among female populations, which
reported inverse associations between dietary protein and MI risk
(4, 5). Dietary protein measured with an FFQ has been validated
against a urinary nitrogen biomarker (41). Our study population
consisted of white male health professionals; thus, it is unclear
whether the results are generalizable to other groups of men.
Conclusions
We observed no association between quintiles of percentage of
energy from dietary protein and risk of total IHD at the protein
levels consumed in the cohort of US men participating in the
Health Professionals Follow-Up Study. We observed a significant
inverse association between higher vegetable protein intake and
risk of fatal IHD. Studies that focus on the health effects of
specific amino acids are warranted.
1271
The authors’ responsibilities were as follows—SRP: study design, data
analysis, and writing of manuscript; MJS: study design, data collection,
and critical revision of manuscript; DS: data analysis and revision of manu-
script; WCW: study design, data collection, and critical revision of manu-
script; and EBR: study design, data collection, and critical revision of
manuscript. There were no conflicts of interest to disclose.
REFERENCES
1. Appel LJ. The effects of protein intake on blood pressure and cardio-
vascular disease. Curr Opin Lipidol 2003;14:55–9.
2. Samaha FF, Iqbal N, Seshadri P, et al. A low-carbohydrate as compared
with a low-fat diet in severe obesity. N Engl J Med 2003;348:2074–81.
3. Wolfe BM, Piche LA. Replacement of carbohydrate by protein in
a conventional-fat diet reduces cholesterol and triglyceride concen-
trations in healthy normolipidemic subjects. Clin Invest Med 1999;22:
140–8.
4. Hu FB, Stampfer MJ, Manson JE, et al. Dietary protein and risk of is-
chemic heart disease in women. Am J Clin Nutr 1999;70:221–7.
Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015
5. Keleman LE, Kushi L, Jacobs DR Jr, Cerhan J. Associations of dietary
protein with disease and mortality in a prospective study of post-
menopausal women. Am J Epidemiol 2005;161:239–49.
6. Mozaffarian D, Ascherio A, Hu FB, et al. Interplay between different
polyunsaturated fatty acids and risk of coronary heart disease in men.
Circulation 2005;111:157–64.
7. Whitney E, Rolfes SR. Understanding nutrition. 10th ed. Belmont, CA:
Thomson Wadsworth, 2005.
8. Willett W. Nutritional epidemiology. 2nd ed. New York, NY: Oxford
University Press, 1998.
9. Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett
WC. Reproducibility and validity of an expanded self-administered
semiquantitative food frequency questionnaire among male health pro-
fessionals. Am J Epidemiol 1992;135:1114–26, discussion 1127–36.
10. Rose GA, Blackburn H. Cardiovascular survey methods. 2nd ed. Ge-
neva, Switzerland: World Health Organization, 1982. (Monograph series
no 58.)
11. Hu FB, Stampfer MJ, Rimm E, et al. Dietary fat and coronary heart
disease: a comparison of approaches for adjusting for total energy intake
and modeling repeated dietary measurements. Am J Epidemiol 1999;
149:531–40.
12. Hu FB. Diet and lifestyle influences on risk of coronary heart disease.
Curr Atheroscler Rep 2009;11:257–63.
13. Mente A, de Koning L, Shannon HS, Anand SS. A systematic review of
the evidence supporting a causal link between dietary factors and cor-
onary heart disease. Arch Intern Med 2009;169:659–69.
14. Nordmann AJ, Nordmann A, Briel M, et al. Effects of low-carbohydrate
versus low-fat diets on weight loss and cardiovascular risk factors. Arch
Intern Med 2006;166:285–93.
15. Virtanen JK, Voutilainen S, Rissanen TH, et al. High dietary methionine
intake increases the risk of acute coronary events in middle-aged men.
Nutr Metab Cardiovasc Dis 2006;16:113–20.
16. Halton TL, Willett WC, Liu S, et al. Low-carbohydrate-diet score and
the risk of coronary heart disease in women. N Engl J Med 2006;355:
1991–2002.
17. Teede HJ, Dalais FS, Kotsopoulos D, Liang YL, Davis S, McGrath BP.
Dietary soy has both beneficial and potentially adverse cardiovascular
effects: a placebo-controlled study in men and postmenopausal women.
J Clin Endocrinol Metab 2001;86:3053–60.
18. Wang YF, Jr WY, Yu D, Champagne C, Appel LJ, Lin PH. The re-
lationship between dietary protein intake and blood pressure: results
from the PREMIER study. J Hum Hypertens 2008;22:745–54.
19. Jenkins DJ, Wong JM, Kendall CW, et al. The effect of a plant-based
low-carbohydrate (“Eco-Atkins”) diet on body weight and blood lipid
concentrations in hyperlipidemic subjects. Arch Intern Med 2009;169:
1046–54.
20. Burke V, Hodgson JM, Beilin LJ, Giangiulioi N, Rogers P, Puddey IB.
Dietary protein and soluble fiber reduce ambulatory blood pressure in
treated hypertensives. Hypertension 2001;38:821–6.
21. Stamler J, Liu K, Ruth KJ, Pryer J, Greenland P. Eight-year blood
pressure change in middle-aged men: relationship to multiple nutrients.
Hypertension 2002;39:1000–6.
22. Elliott P, Stamler J, Dyer AR, et al. Association between protein intake
and blood pressure: the INTERMAP Study. Arch Intern Med 2006;166:
79–87.
1272 PREIS ET AL
23. Anderson JW, Johnstone BM, Cook-Newell ME. Meta-analysis of the
effects of soy protein intake on serum lipids. N Engl J Med 1995;333:
276–82.
24. Sacks FM, Lichtenstein A, Van Horn L, Harris W, Kris-Etherton P,
Winston M. Soy protein, isoflavones, and cardiovascular health: an
American Heart Association Science Advisory for professionals from
the Nutrition Committee. Circulation 2006;113:1034–44.
25. Stern L, Iqbal N, Seshadri P, et al. The effects of low-carbohydrate
versus conventional weight loss diets in severely obese adults: one
-year follow-up of a randomized trial. Ann Intern Med 2004;140:
778–85.
26. Foster GD, Wyatt H, Hill J, et al. A randomized trial of a low-
carbohydrate diet for obesity. N Engl J Med 2003;348:2082–90.
27. Yancy WS, Olsen MK, Guyton JR, Bakst RP, Westman EC. A low-
carbohydrate, ketogenic diet versus a low-fat diet to treat obesity and
hyperlipidemia: a randomized, controlled trial. Ann Intern Med 2004;
140:769–77.
28. Dansinger ML, Gleason JA, Griffith JL, Selker HP, Schaefer EJ. Com-
parison of the Atkins, Ornish, Weight Watchers, and Zone diets for
weight loss and heart disease risk reduction. JAMA 2005;293:43–53.
29. Clifton PM, Bastiaans K, Keogh JB. High protein diets decrease total
and abdominal fat and improve CVD risk profile in overweight and
obese men and women with elevated triacylglycerol. Nutr Metab Car-
diovasc Dis 2009;19:548–54.
30. Brehm BJ, Seeley R, Daniels S, D’Alessio D. A randomized trial
comparing a very low carbohydrate diet and a calorie-restricted low fat
diet on body weight and cardiovascular risk factors in healthy women.
J Clin Endocrinol Metab 2003;88:1617–23.
31. Nickols-Richardson SM, Coleman MD, Volpe JJ, Hosig KW. Perceived
hunger is lower and weight loss is greater in overweight premenopausal
women consuming a low-carbohydrate/high-protein vs high-
carbohydrate/low-fat diet. J Am Diet Assoc 2005;105:1433–7.
32. Johnstone AM, Horgan G, Murison S, Bremner D, Lobley G. Effects of
a high-protein ketogenic diet on hunger, appetite, and weight loss in
obese men feeding ad libitum. Am J Clin Nutr 2008;87:44–55.
33. Shai I, Schwarzfuchs D, Henkin Y, et al. Weight loss with a low-
carbohydrate, Mediterranean, or low-fat diet. N Engl J Med 2008;359:
229–41.
34. Vetter ML, Iqbal N, Dalton-Bakes C, Volger S, Wadden TA. Long-term
effects of low-carbohydrate versus low-fat diets in obese persons. Ann
Intern Med 2010;152:334–5.
35. Hu FB, Stampfer MJ, Manson JE, et al. Frequent nut consumption and
risk of coronary heart disease in women: prospective cohort study. BMJ
1998;317:1341–5.
36. Bazzano LA, He J, Ogden LG, et al. Legume consumption and risk of
coronary heart disease in US men and women: NHANES I Epidemio-
logic Follow-up Study. Arch Intern Med 2002;161:2573–8.
37. Albert CM, Oh K, Whang W, et al. Dietary alpha-linolenic acid intake
and risk of sudden cardiac death and coronary heart disease. Circulation
2005;112:3232–8.
38. Krajcovicova-Kudlackova M, Babinska K, Valachovicova M. Health
benefits and risks of plant proteins. Bratisl Lek Listy2005;106:231–4.
Downloaded from ajcn.nutrition.org at BRESCIA UNIVERSITY COLLEGE on March 8, 2015
39. Obarzanek E, Velletri PA, Cutler JA. Dietary protein and blood pressure.
JAMA 1996;275:1598–603.
40. Stamler J, Brown IJ, Daviglus ML, et al. Glutamic acid, the main dietary
amino acid, and blood pressure: the INTERMAP Study (International
Collaborative Study of Macronutrients, Micronutrients and Blood
Pressure). Circulation 2009;120:221–8.
41. Kroke A, Klipstein-Grobusch K, Voss S, et al. Validation of a self-
administered food-frequency questionnaire administered in the Euro-
pean Prospective Investigation into Cancer and Nutrition (EPIC) Study:
comparison of energy, protein, and macronutrient intakes estimated with
the doubly labeled water, urinary nitrogen, and repeated 24-h dietary
recall methods. Am J Clin Nutr 1999;70:439–47.