Red Meat Consumption and MortalityResults From 2 Prospective Cohort Studies

Abstract

Red meat consumption has been associated with an increased risk of chronic diseases. However, its relationship with mortality remains uncertain.
We prospectively observed 37 698 men from the Health Professionals Follow-up Study (1986-2008) and 83 644 women from the Nurses' Health Study (1980-2008) who were free of cardiovascular disease (CVD) and cancer at baseline. Diet was assessed by validated food frequency questionnaires and updated every 4 years.
We documented 23 926 deaths (including 5910 CVD and 9464 cancer deaths) during 2.96 million person-years of follow-up. After multivariate adjustment for major lifestyle and dietary risk factors, the pooled hazard ratio (HR) (95% CI) of total mortality for a 1-serving-per-day increase was 1.13 (1.07-1.20) for unprocessed red meat and 1.20 (1.15-1.24) for processed red meat. The corresponding HRs (95% CIs) were 1.18 (1.13-1.23) and 1.21 (1.13-1.31) for CVD mortality and 1.10 (1.06-1.14) and 1.16 (1.09-1.23) for cancer mortality. We estimated that substitutions of 1 serving per day of other foods (including fish, poultry, nuts, legumes, low-fat dairy, and whole grains) for 1 serving per day of red meat were associated with a 7% to 19% lower mortality risk. We also estimated that 9.3% of deaths in men and 7.6% in women in these cohorts could be prevented at the end of follow-up if all the individuals consumed fewer than 0.5 servings per day (approximately 42 g/d) of red meat.
Red meat consumption is associated with an increased risk of total, CVD, and cancer mortality. Substitution of other healthy protein sources for red meat is associated with a lower mortality risk.

Full text

ONLINE FIRST
ORIGINAL INVESTIGATION
Red Meat Consumption and Mortality
Results From 2 Prospective Cohort Studies
An Pan, PhD; Qi Sun, MD, ScD; Adam M. Bernstein, MD, ScD; Matthias B. Schulze, DrPH;
JoAnn E. Manson, MD, DrPH; Meir J. Stampfer, MD, DrPH; Walter C. Willett, MD, DrPH; Frank B. Hu, MD, PhD
Background:Red meat consumption has been associ-
ated with an increased risk of chronic diseases. How-
ever, its relationship with mortality remains uncertain.
Methods:We prospectively observed 37 698 men from
the Health Professionals Follow-up Study (1986-2008)
and 83 644 women from the Nurses’ Health Study (1980-
2008) who were free of cardiovascular disease (CVD) and
cancer at baseline. Diet was assessed by validated food
frequency questionnaires and updated every 4 years.
Results:We documented 23 926 deaths (including 5910
CVD and 9464 cancer deaths) during 2.96 million person-
years of follow-up. After multivariate adjustment for ma-
jor lifestyle and dietary risk factors, the pooled hazard
ratio (HR) (95% CI) of total mortality for a 1-serving-
per-day increase was 1.13 (1.07-1.20) for unprocessed
red meat and 1.20 (1.15-1.24) for processed red meat.
The corresponding HRs (95% CIs) were 1.18 (1.13-
1.23) and 1.21 (1.13-1.31) for CVD mortality and 1.10
(1.06-1.14) and 1.16 (1.09-1.23) for cancer mortality. We
estimated that substitutions of 1 serving per day of other
foods (including fish, poultry, nuts, legumes, low-fat dairy,
and whole grains) for 1 serving per day of red meat were
associated with a 7% to 19% lower mortality risk. We also
estimated that 9.3% of deaths in men and 7.6% in women
in these cohorts could be prevented at the end of fol-
low-up if all the individuals consumed fewer than 0.5 serv-
ings per day (approximately 42 g/d) of red meat.
Conclusions:Red meat consumption is associated with
an increased risk of total, CVD, and cancer mortality. Sub-
stitution of other healthy protein sources for red meat is
associated with a lower mortality risk.
Arch Intern Med.
Published online March 12, 2012.
doi:10.1001/archinternmed.2011.2287
MEAT IS A MAJOR SOURCE
of protein and fat in
most diets. Substantial
evidence from epide-
miological studies
shows that consumption of meat, particularly
red meat, is associated with increased risks
of diabetes,1cardiovascular disease (CVD),2
and certain cancers.3Several studies also sug-
gest an elevated risk of mortality associated
with red meat intake. However, most of these
studies have been performed in populations
with a particularly high proportion of veg-
etarians (such as Seventh-Day Adventists in
the United States4and several studies in
Europe5). A recent large cohort study6with
10 years of follow-up found that a higher in-
take of total red meat and total processed meat
was associated with an increased risk of mor-
tality. However, this study did not differen-
tiate unprocessed from processed red meat,
and diet and other covariates were assessed
at baseline only. Furthermore, to our knowl-
edge, no study has examined whether sub-
stitution of other dietary components for red
meat is associated with a reduced mortality
risk.
Therefore, we investigated the associa-
tion between red meat intake and cause-
specific and total mortality in 2 large co-
horts with repeated measures of diet and
up to 28 years of follow-up: the Health Pro-
fessionals Follow-up Study (HPFS) and the
Nurses’ Health Study (NHS). We also es-
timated the associations of substituting
other healthy protein sources for red meat
with total and cause-specific mortality.
METHODS
STUDY POPULATION
We analyzed data from 2 prospective cohort stud-
ies: the HPFS (initiated in 1986, n=51 529 men
aged 40-75 years) and the NHS (started in 1976,
n=121 700 women aged 30-55 years). Detailed
descriptions of the cohorts are provided else-
where.7,8 Questionnaires were administered bi-
ennially to collect and update medical, lifestyle,
and other health-related information, and the fol-
low-up rates exceeded 90% in each 2-year cycle
for both cohorts.
See Invited Commentary
at end of article
Author Affiliations:
Departments of Nutrition
(Drs Pan, Sun, Bernstein,
Stampfer, Willett, and Hu) and
Epidemiology (Drs Manson,
Stampfer, Willett, and Hu),
Harvard School of Public
Health, and Channing
Laboratory (Drs Sun, Stampfer,
Willett, and Hu) and Division
of Preventive Medicine
(Dr Manson), Department of
Medicine, Brigham and
Women’s Hospital and Harvard
Medical School, Boston,
Massachusetts; Wellness
Institute of the Cleveland
Clinic, Lyndhurst, Ohio
(Dr Bernstein); and Department
of Molecular Epidemiology,
German Institute of Human
Nutrition, Nuthetal, Germany
(Dr Schulze).
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E1
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

In the present analysis, we used 1986 for the HPFS and 1980
for the NHS as baseline, when we assessed diet using a validated
food frequency questionnaire (FFQ); 49 934 men and 92 468 wom-
en returned the baseline FFQ. We excluded 5617 men and 5613
women who had a history of CVD or cancer at baseline and 6619
men and 3211 women who left more than 9 blank responses on
the baseline FFQ, had missing information about meat intake, or
reported implausible energy intake levels (⬍500 or ⬎3500 kcal/
d). After the exclusions, data from 37 698 men and 83 644 wom-
en were available for the analysis. The excluded participants and
those who remained in the study were similar with respect to red
meat intake and obesity status at baseline. The study protocol was
approved by the institutional review boards of Brigham and Wom-
en’s Hospital and Harvard School of Public Health.
ASSESSMENT OF MEAT CONSUMPTION
In 1980, a 61-item FFQ was administered to the NHS participants
to collect information about their usual intake of foods and bev-
erages in the previous year. In 1984, 1986, 1990, 1994, 1998, 2002,
and 2006, similar but expanded FFQs with 131 to 166 items were
sent to these participants to update their diet. Using the expanded
FFQ used in the NHS, dietary data were collected in 1986, 1990,
1994, 1998, 2002, and 2006 from the HPFS participants. In each
FFQ, we asked the participants how often, on average, they con-
sumed each food of a standard portion size. There were 9 possible
responses, ranging from “never or less than once per month” to
“6 or more times per day.” Questionnaire items about unprocessed
red meat consumption included “beef, pork, or lamb as main dish”
(pork was queried separately beginning in 1990), “hamburger,”
and “beef, pork, or lamb as a sandwich or mixed dish.” The stan-
dard serving size was 85 g (3 oz) for unprocessed red meat. Pro-
cessed red meat included “bacon” (2 slices, 13 g), “hot dogs” (one,
45 g), and “sausage, salami, bologna, and other processed red meats”
(1 piece, 28 g). The reproducibility and validity of these FFQs have
been described in detail elsewhere.9,10 The corrected correlation
coefficients between the FFQ and multiple dietary records were
0.59 for unprocessed red meat and 0.52 for processed red meat in
the HPFS,9and similar correlations were found in the NHS.10
ASCERTAINMENT OF DEATH
The ascertainment of death has been documented in previous stud-
ies.11 Briefly, deaths were identified by reports from next of kin,
via postal authorities, or by searching the National Death Index,
and at least 95% of deaths were identified.11 The cause of death
was determined after review by physicians and were primarily
based on medical records and death certificates. We used the In-
ternational Classification of Diseases, Eighth Revision, which was
widely used at the start of the cohorts, to distinguish deaths due
to cancer (codes 140-207) and CVDs (codes 390-459 and 795).
ASSESSMENT OF COVARIATES
In the biennial follow-up questionnaires, we inquired and up-
dated information on medical, lifestyle, and other health-
related factors, such as body weight; cigarette smoking status;
physical activity level; medication or supplement use; family
history of diabetes mellitus, myocardial infarction, and can-
cer; and history of diabetes mellitus, hypertension, and hyper-
cholesterolemia. In NHS participants, we also ascertained meno-
pausal status and postmenopausal hormone use.
STATISTICAL ANALYSIS
We used time-dependent Cox proportional hazards regres-
sion models to assess the association of red meat consumption
with cause-specific and total mortality risks during follow-up.
We conducted analyses separately for each cohort. In multi-
variate analysis, we simultaneously controlled for intakes of total
energy, whole grains, fruits, and vegetables (all in quintiles)
and for other potential nondietary confounding variables with
updated information at each 2- or 4-year questionnaire cycle.
These variables included age; body mass index (calculated as
weight in kilograms divided by height in meters squared)
(⬍23.0, 23.0-24.9, 25.0-29.9, 30.0-34.9, or ⱖ35.0); race (white
or nonwhite); smoking status (never, past, or current [1-14,
15-24, or ⱖ25 cigarettes per day]); alcohol intake (0, 0.1-4.9,
5.0-14.9, or ⱖ15.0 g/d in women; 0, 0.1-4.9, 5.0-29.9, or ⱖ30.0
g/d in men); physical activity level (⬍3.0, 3.0-8.9, 9.0-17.9, 18.0-
26.9, or ⱖ27.0 hours of metabolic equivalent tasks per week);
multivitamin use (yes or no); aspirin use (yes or no); family
history of diabetes mellitus, myocardial infarction, or cancer;
and baseline history of diabetes mellitus, hypertension, or hy-
percholesterolemia. In women, we also adjusted for postmeno-
pausal status and menopausal hormone use.
To better represent long-term diet and to minimize within-
person variation, we created cumulative averages of food in-
take from baseline to death from the repeated FFQs.12 We re-
placed missing values in each follow-up FFQ with the cumulative
averages before the missing values. We stopped updating the
dietary variables when the participants reported a diagnosis of
diabetes mellitus, stroke, coronary heart disease, angina, or can-
cer because these conditions might lead to changes in diet.
We conducted several sensitivity analyses to test the ro-
bustness of the results: (1) we further adjusted for intakes of
other major dietary variables (fish, poultry, nuts, legumes, and
dairy products, all in quintiles) or several nutrients or dietary
components (glycemic load, cereal fiber, magnesium, and poly-
unsaturated and trans fatty acids, all in quintiles) instead of foods;
(2) we corrected for measurement error13 in the assessment of
red meat intake by using a regression calibration approach using
data from validation studies conducted in the HPFS9in 1986
and in the NHS10 in 1980 and 1986; (3) we repeated the analy-
sis by using simply updated dietary methods (using the most
recent dietary variables to predict mortality risk in the next 4
years)12 or continue to update a participant’s diet even after he
or she reported a diagnosis of major chronic disease or using
only baseline dietary variables; and (4) we used the energy
density of red meat intake (serving/1000 kcal⫻d−1) as the ex-
posure instead of the crude intake. In addition, we used re-
stricted cubic spline regressions with 4 knots to examine a dose-
response relation between red meat intake and risk of total
mortality.
We estimated the associations of substituting 1 serving of
an alternative food for red meat with mortality by including
both as continuous variables in the same multivariate model,
which also contained nondietary covariates and total energy
intake. The difference in their ␤coefficients and in their own
variances and covariance were used to estimate the hazard ra-
tios (HRs) and 95% CIs for the substitution associations.14 We
calculated population-attributable risk (95% CI) to estimate the
proportion of deaths in the 2 cohorts that would be prevented
at the end of follow-up if all the participants were in the low-
intake group.15 For these analyses, we compared participants
in the low–red meat intake category (⬍0.5 servings daily, or
42 g/d) with the remaining participants in the cohorts.
The HRs from the final multivariate-adjusted models in each
cohort were pooled to obtain a summary risk estimate with the
use of an inverse variance–weighted meta-analysis by the ran-
dom-effects model, which allowed for between-study hetero-
geneity. Data were analyzed using a commercially available soft-
ware program (SAS, version 9.2; SAS Institute, Inc), and
statistical significance was set at a 2-tailed ␣=.05.
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E2
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

RESULTS
In the HPFS, with up to 22 years of follow-up (758 524 per-
son-years), we documented 8926 deaths, of which 2716
were CVD deaths and 3073 were cancer deaths. In the NHS,
with up to 28 years of follow-up (2 199 892 person-
years), we documented 15 000 deaths, of which 3194 were
CVD deaths and 6391 were cancer deaths. For both co-
horts combined, we documented 23 926 deaths (includ-
ing 5910 CVD deaths and 9464 cancer deaths) during 2.96
million person-years of follow-up. Men and women with
higher intake of red meat were less likely to be physically
active and were more likely to be current smokers, to drink
alcohol, and to have a higher body mass index (Table 1).
Table 1. Baseline Age-Standardized Characteristics of Participants in the 2 Cohorts According to Quintiles
of Total Red Meat Consumption
Characteristic
Total Red Meat Intake Quintile, Servings per Day
Q1 Q2 Q3 Q4 Q5
Health Professionals Follow-up Study
Participants, No. 7431 7813 7308 7606 7540
Age, mean, y 53.8 52.6 52.5 52.5 52.2
Total red meat intake, mean, servings per day 0.22 0.62 1.01 1.47 2.36
Physical activity, mean, MET-h/wk 27.5 22.7 20.2 18.8 17.2
Body mass index, meana24.7 25.3 25.5 25.7 26.0
White race, % 93.1 95.1 95.2 95.8 95.8
Current smoker, % 5.0 7.3 9.8 11.3 14.5
Diabetes mellitus, % 2.0 2.0 2.2 2.4 3.5
Hypertension, % 19.5 19.7 19.3 19.6 20.2
High cholesterol, % 14.8 11.1 9.7 9.0 7.9
Family history of diabetes mellitus, % 19.5 18.6 19.1 20.0 19.3
Family history of myocardial infarction, % 35.1 31.8 30.9 31.4 30.0
Family history of cancer, % 33.7 34.5 35.0 33.9 33.6
Current multivitamin use, % 49.1 42.5 40.3 39.5 36.6
Current aspirin use, % 24.6 26.4 25.9 27.8 27.4
Dietary intake, mean
Total energy, kcal/d 1659 1752 1886 2091 2396
Alcohol, g/d 8.4 10.7 11.2 12.4 13.4
Fruit, servings per day 2.83 2.35 2.21 2.13 2.04
Vegetables, servings per day 3.29 2.89 2.91 2.97 3.07
Whole grains, servings per day 1.93 1.58 1.50 1.51 1.48
Nuts, servings per day 0.45 0.45 0.44 0.47 0.49
Legumes, servings per day 0.45 0.38 0.39 0.43 0.47
Dairy products, servings per day 1.65 1.80 1.89 2.02 2.14
Fish, servings per day 0.55 0.43 0.38 0.36 0.32
Poultry, servings per day 0.64 0.58 0.55 0.55 0.53
Nurses’ Health Study
Participants, No. 16 499 17 247 16 461 16 603 16 834
Age, mean, y 47.3 46.0 45.8 45.3 46.0
Total red meat intake, mean, servings per day 0.53 1.04 1.52 2.01 3.10
Physical activity, mean, MET-h/wk 16.9 13.9 13.8 13.3 12.4
Body mass index, meana23.9 24.3 24.4 24.5 24.7
White race, % 96.9 97.9 97.8 98.0 97.2
Current smoker, % 25.5 29.1 28.2 29.7 31.6
Diabetes mellitus, % 1.6 1.8 2.1 2.2 2.9
Hypertension, % 15.2 15.7 15.5 15.4 16.4
High cholesterol, % 6.0 5.3 5.2 4.5 4.7
Family history of diabetes, % 26.7 27.9 28.1 29.0 29.9
Family history of myocardial infarction, % 19.4 19.0 19.0 18.6 19.0
Family history of cancer, % 17.1 16.7 16.1 16.6 16.3
Postmenopausal, % 31.3 31.3 30.8 31.1 31.1
Current menopausal hormone use, %b20.6 20.4 21.0 21.3 20.7
Current multivitamin use, % 37.9 33.6 33.1 32.8 32.3
Current aspirin use, % 43.2 46.9 46.3 48.3 49.1
Dietary intake, mean
Total energy, kcal/d 1202 1371 1523 1705 2030
Alcohol, g/d 5.8 6.3 6.6 6.5 6.6
Fruit, servings per day 2.21 2.05 2.04 2.03 2.02
Vegetables, servings per day 1.89 1.83 1.92 1.98 2.08
Whole grains, servings per day 1.53 1.37 1.35 1.36 1.28
Nuts, servings per day 0.16 0.13 0.13 0.14 0.15
Legumes, servings per day 0.44 0.44 0.45 0.49 0.52
Dairy products, servings per day 1.81 1.80 1.82 1.87 1.83
Fish, servings per day 0.50 0.40 0.39 0.35 0.33
Poultry, servings per day 0.64 0.59 0.59 0.58 0.58
Abbreviation: MET-h, hours of metabolic equivalent tasks.
aBody mass index is calculated as weight in kilograms divided by height in meters squared.
bCurrent menopausal hormone use in postmenopausal women.
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E3
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

In addition, a higher red meat intake was associated with
a higher intake of total energy but lower intakes of whole
grains, fruits, and vegetables. Unprocessed and processed
red meat consumption was moderately correlated (r=0.40
in the HPFS and 0.37 in the NHS). However, red meat con-
sumption was less correlated with intakes of poultry and
fish (Spearman correlation coefficients, r=−0.04 and −0.18
in the HPFS and r=0.05 and −0.12 in the NHS, respec-
tively). During follow-up, red meat intake declined in men
and women (eFigure; http://www.archinternmed.com). For
example, the mean daily intake of unprocessed red meat
dropped from 0.75 to 0.63 servings from 1986 to 2006 in
men and from 1.10 to 0.55 servings from 1980 to 2006 in
women.
Unprocessed and processed red meat intakes were as-
sociated with an increased risk of total, CVD, and cancer
mortality in men and women in the age-adjusted and fully
adjusted models (Tables 2,3, and 4). When treating red
meat intake as a continuous variable, the elevated risk of
total mortality in the pooled analysis for a 1-serving-per-
day increase was 12% (HR, 1.12; 95% CI, 1.09-1.15) for
total red meat, 13% (HR, 1.13; 95% CI, 1.07-1.20) for un-
processed red meat, and 20% (HR, 1.20; 95% CI, 1.15-
1.24) for processed red meat. The HRs (95% CIs) for CVD
mortality were 1.16 (1.12-1.20) for total red meat, 1.18
(1.13-1.23) for unprocessed red meat, and 1.21 (1.13-
1.31) for processed red meat. The HRs (95% CIs) for can-
cer mortality were 1.10 (1.07-1.13) for total red meat, 1.10
(1.06-1.14) for unprocessed red meat, and 1.16 (1.09-
1.23) for processed red meat. We found no statistically sig-
nificant differences among specific unprocessed red meat
items or among specific processed red meat items for the
associations with total mortality (eTable 1). However, ba-
con and hot dogs tended to be associated with a higher risk
than other items. Spline regression analysis showed that
the association between red meat intake and risk of total
Table 2. All-Cause Mortality According to Red Meat Intake in the Health Professionals Follow-up Study and the Nurses’ Health Study
Variable
Frequency of Consumption Quintilesa
PValue
for Trend
HR (95% CI) for a
1-Serving-per-Day
IncreaseQ1 Q2 Q3 Q4 Q5
Health Professionals Follow-up Study
Total red meat, servings
per dayb0.25 (0.13-0.37) 0.61 (0.53-0.70) 0.95 (0.87-1.04) 1.36 (1.24-1.49) 2.07 (1.83-2.47) NA NA
Cases/person-years, No. 1713/151 212 1610/152 120 1679/151 558 1794/152318 2130/151 315 NA NA
Age-adjusted model 1 [Reference] 1.06 (0.99-1.14) 1.14 (1.06-1.21) 1.21 (1.14-1.30) 1.45 (1.36-1.54) ⬍.001 1.19 (1.16-1.23)
Multivariate modelc1 [Reference] 1.12 (1.05-1.20) 1.21 (1.13-1.30) 1.25 (1.16-1.34) 1.37 (1.27-1.47) ⬍.001 1.14 (1.10-1.17)
Unprocessed red meat,
servings per dayb0.17 (0.07-0.24) 0.43 (0.37-0.47) 0.66 (0.58-0.73) 0.95 (0.87-1.04) 1.46 (1.29-1.67) NA NA
Cases/person-years, No. 1855/150 676 1722/149 097 1535/154 352 1819/150 925 1995/153 474 NA NA
Age-adjusted model 1 [Reference] 1.06 (0.99-1.13) 1.00 (0.94-1.07) 1.15 (1.08-1.23) 1.34 (1.25-1.42) ⬍.001 1.22 (1.18-1.27)
Multivariate modelc1 [Reference] 1.11 (1.04-1.18) 1.14 (1.06-1.22) 1.20 (1.12-1.28) 1.29 (1.20-1.38) ⬍.001 1.17 (1.12-1.21)
Processed red meat,
servings per dayb0.02 (0-0.07) 0.13 (0.10-0.14) 0.21 (0.20-0.26) 0.39 (0.34-0.46) 0.74 (0.64-1.00) NA NA
Cases/person-years, No. 1917/171 619 1395/131 069 1661/152 481 1717/152 128 2236/151 227 NA NA
Age-adjusted model 1 [Reference] 0.99 (0.93-1.06) 1.13 (1.05-1.20) 1.14 (1.07-1.22) 1.38 (1.30-1.47) ⬍.001 1.34 (1.28-1.40)
Multivariate modelc1 [Reference] 1.06 (0.99-1.14) 1.15 (1.07-1.23) 1.18 (1.10-1.27) 1.27 (1.19-1.36) ⬍.001 1.18 (1.12-1.24)
Nurses’ Health Study
Total red meat, servings
per dayb0.51 (0.37-0.61) 0.85 (0.76-0.96) 1.14 (1.03-1.32) 1.49 (1.33-1.71) 2.17 (1.85-2.66) NA NA
Cases/person-years, No. 2946/438 326 2759/442 134 2658/439 712 2872/440 329 3765/439 391 NA NA
Age-adjusted model 1 [Reference] 1.07 (1.01-1.12) 1.09 (1.04-1.15) 1.24 (1.18-1.30) 1.61 (1.53-1.69) ⬍.001 1.30 (1.28-1.33)
Multivariate modelc1 [Reference] 1.08 (1.02-1.14) 1.11 (1.05-1.17) 1.18 (1.12-1.24) 1.24 (1.17-1.30) ⬍.001 1.11 (1.08-1.13)
Unprocessed red meat,
servings per dayb0.37 (0.28-0.46) 0.61 (0.56-0.68) 0.86 (0.77-1.00) 1.13 (1.01-1.28) 1.64 (1.43-2.05) NA NA
Cases/person-years, No. 3079/441 041 2885/441 207 2545/439 306 2709/431 097 3782/447 240 NA NA
Age-adjusted model 1 [Reference] 1.05 (1.00-1.11) 0.98 (0.93-1.03) 1.09 (1.03-1.14) 1.48 (1.41-1.55) ⬍.001 1.31 (1.28-1.35)
Multivariate modelc1 [Reference] 1.07 (1.01-1.12) 1.07 (1.01-1.12) 1.10 (1.05-1.16) 1.19 (1.13-1.25) ⬍.001 1.10 (1.06-1.13)
Processed red meat,
servings, per dayb0.05 (0-0.11) 0.14 (0.13-0.16) 0.23 (0.21-0.28) 0.36 (0.33-0.42) 0.64 (0.56-0.87) NA NA
Cases/person-years, No. 3076/442 594 2799/420 403 2778/455 365 2814/441 369 3533/440 161 NA NA
Age-adjusted model 1 [Reference] 1.06 (1.01-1.12) 1.10 (1.04-1.16) 1.18 (1.12-1.24) 1.49 (1.42-1.56) ⬍.001 1.61 (1.54-1.69)
Multivariate modelc1 [Reference] 1.04 (0.99-1.10) 1.08 (1.03-1.14) 1.14 (1.08-1.20) 1.20 (1.14-1.27) ⬍.001 1.21 (1.15-1.27)
Pooled Resultsd
Total red meat 1 [Reference] 1.10 (1.05-1.14) 1.15 (1.06-1.26) 1.21 (1.14-1.28) 1.30 (1.18-1.43) ⬍.001 1.12 (1.09-1.15)
Unprocessed red meat 1 [Reference] 1.08 (1.05-1.12) 1.10 (1.03-1.17) 1.15 (1.05-1.25) 1.23 (1.14-1.34) ⬍.001 1.13 (1.07-1.20)
Processed red meat 1 [Reference] 1.05 (1.00-1.09) 1.11 (1.04-1.18) 1.15 (1.11-1.20) 1.23 (1.16-1.30) ⬍.001 1.20 (1.15-1.24)
Abbreviations: HR, hazard ratio; NA, not applicable.
aData are given as HR (95% CI) except where indicated otherwise.
bData are given as median (interquartile range).
cThe multivariate model was adjusted for age (continuous); body mass index (calculated as weight in kilograms divided by height in meters squared) category
(⬍23.0, 23.0-24.9, 25.0-29.9, 30.0-34.9, or ⱖ35.0); alcohol consumption (0, 0.1-4.9, 5.0-29.9, or ⱖ30.0 g/d in men; 0, 0.1-4.9, 5.0-14.9, or ⱖ15.0 g/d in
women); physical activity level (⬍3.0, 3.0-8.9, 9.0-17.9, 18.0-26.9, or ⱖ27.0 hours of metabolic equivalent tasks per week); smoking status (never, past, or
current [1-14, 15-24, or ⱖ25 cigarettes per day]); race (white or nonwhite); menopausal status and hormone use in women (premenopausal, postmenopausal
never users, postmenopausal past users, or postmenopausal current users); family history of diabetes mellitus, myocardial infarction, or cancer; history of
diabetes mellitus, hypertension, or hypercholesterolemia; and intakes of total energy, whole grains, fruits, and vegetables, all in quintiles.
dResults from the multivariate model were combined using the random-effects model.
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E4
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

mortality was linear (P⬍.001 for linearity; Figure 1). Fur-
thermore, no significant interaction was detected be-
tween red meat intake and body mass index or physical ac-
tivity level (P⬎.10 for both tests).
Additional adjustment for other foods (fish, poultry, nuts,
beans, and dairy products) or nutrients (glycemic load, ce-
real fiber, magnesium, and polyunsaturated and trans fatty
acids) did not appreciably alter the results. Additional ad-
justment for saturated fat and cholesterol moderately at-
tenuated the association between red meat intake and risk
of CVD death, and the pooled HR (95% CI) dropped from
1.16 (1.12-1.20) to 1.12 (1.07-1.18). Similarly, additional
adjustment for heme iron moderately attenuated the asso-
ciation, and the pooled HR (95% CI) dropped from 1.16
(1.12-1.20) to 1.11 (1.05-1.17). Additional adjustment for
husband’s educational level as a surrogate of socioeco-
nomic status in women did not change the results.
The results were not materially changed when we con-
tinued to update dietary information even after the di-
agnosis of chronic diseases (eTable 2) or simply up-
dated the dietary variables (eTable 3). Also, using the
energy density of red meat intake as the exposure showed
similar findings (eTable 4). In the sensitivity analysis that
accounted for measurement error in diet, the associa-
tions became even stronger. For example, the HR was 1.25
(95% CI, 1.16-1.35) for a 1-serving-per-day increase in
total red meat intake with mortality in the HPFS, and it
was 1.83 (95% CI, 1.54-2.20) in the NHS. However, the
associations were attenuated in analyses using only base-
line dietary data (eTable 5).
In the substitution analyses, replacing 1 serving of
total red meat with 1 serving of fish, poultry, nuts,
legumes, low-fat dairy products, or whole grains daily
was associated with a lower risk of total mortality: 7%
(HR, 0.93; 95% CI, 0.90-0.97) for fish, 14% (HR, 0.86;
95% CI, 0.82-0.91) for poultry, 19% (HR, 0.81; 95% CI,
0.77-0.86) for nuts, 10% (HR, 0.90; 95% CI, 0.86-0.94)
for legumes, 10% (HR, 0.90; 95% CI, 0.86-0.94) for
low-fat dairy products, and 14% (HR, 0.86; 95% CI,
0.82-0.88) for whole grains (Figure 2). The corre-
sponding substitution estimates were 5%, 13%, 18%,
8%, 9%, and 13% for replacement of unprocessed red
meat and 10%, 17%, 22%, 13%, 13%, and 16% for
replacement of processed red meat.
Table 3. Cardiovascular Mortality According to Red Meat Intake in the Health Professionals Follow-up Study
and the Nurses’ Health Study
Variable
Frequency of Consumption Quintilesa
PValue
for Trend
HR (95% CI) for a
1-Serving-per-Day
IncreaseQ1 Q2 Q3 Q4 Q5
Health Professionals Follow-up Study
Total red meat
Cases/person-years, No. 537/152 293 490/153 126 506/152 623 518/153 454 665/152 647 NA NA
Age-adjusted model 1 [Reference] 1.05 (0.93-1.19) 1.11 (0.98-1.26) 1.15 (1.02-1.30) 1.48 (1.32-1.66) ⬍.001 1.21 (1.16-1.27)
Multivariate modelb1 [Reference] 1.09 (0.96-1.24) 1.16 (1.03-1.32) 1.17 (1.03-1.33) 1.35 (1.19-1.53) ⬍.001 1.14 (1.08-1.20)
Unprocessed red meat
Cases/person-years, No. 578/151 850 528/150 172 446/155 316 532/152 087 632/154 719 NA NA
Age-adjusted model 1 [Reference] 1.08 (0.95-1.20) 0.97 (0.86-1.10) 1.11 (0.98-1.25) 1.41 (1.26-1.58) ⬍.001 1.26 (1.18-1.34)
Multivariate modelb1 [Reference] 1.10 (0.97-1.24) 1.08 (0.95-1.22) 1.14 (1.01-1.29) 1.32 (1.16-1.49) ⬍.001 1.19 (1.10-1.27)
Processed red meat
Cases/person-years, No. 594/172 817 423/131 953 510/153 537 512/153 206 677/152 631 NA NA
Age-adjusted model 1 [Reference] 0.99 (0.88-1.12) 1.14 (1.01-1.29) 1.13 (1.00-1.27) 1.37 (1.23-1.53) ⬍.001 1.34 (1.24-1.46)
Multivariate modelb1 [Reference] 1.05 (0.93-1.19) 1.15 (1.01-1.30) 1.15 (1.02-1.31) 1.25 (1.11-1.41) .003 1.17 (1.07-1.29)
Nurses’ Health Study
Total red meat
Cases/person-years, No. 601/440 429 570/444 046 517/441 619 598/442 319 908/441 994 NA NA
Age-adjusted model 1 [Reference] 1.11 (0.99-1.25) 1.09 (0.97-1.22) 1.33 (1.19-1.49) 1.98 (1.79-2.20) ⬍.001 1.44 (1.38-1.50)
Multivariate modelb1 [Reference] 1.14 (1.01-1.27) 1.11 (0.99-1.26) 1.28 (1.13-1.43) 1.45 (1.30-1.63) ⬍.001 1.17 (1.11-1.22)
Unprocessed red meat
Cases/person-years, No. 617/443 224 646/443 182 481/441 163 549/432 988 901/449 850 NA NA
Age-adjusted model 1 [Reference] 1.21 (1.08-1.35) 0.96 (0.85-1.09) 1.15 (1.03-1.29) 1.82 (1.65-2.02) ⬍.001 1.46 (1.39-1.54)
Multivariate modelb1 [Reference] 1.22 (1.09-1.37) 1.09 (0.96-1.23) 1.19 (1.06-1.34) 1.39 (1.24-1.55) ⬍.001 1.17 (1.10-1.24)
Processed red meat
Cases/person-years, No. 671/444 737 551/422 411 586/457 265 572/443 383 814/442 609 NA NA
Age-adjusted model 1 [Reference] 0.98 (0.88-1.10) 1.10 (0.99-1.23) 1.16 (1.03-1.29) 1.65 (1.49-1.83) ⬍.001 1.79 (1.64-1.95)
Multivariate modelb1 [Reference] 0.97 (0.87-1.09) 1.10 (0.99-1.23) 1.12 (0.99-1.25) 1.29 (1.15-1.43) ⬍.001 1.26 (1.15-1.39)
Pooled Resultsc
Total red meat 1 [Reference] 1.12 (1.03-1.22) 1.13 (1.04-1.24) 1.23 (1.13-1.34) 1.40 (1.29-1.53) ⬍.001 1.16 (1.12-1.20)
Unprocessed red meat 1 [Reference] 1.16 (1.05-1.28) 1.09 (1.00-1.18) 1.17 (1.07-1.27) 1.36 (1.25-1.47) ⬍.001 1.18 (1.13-1.23)
Processed red meat 1 [Reference] 1.01 (0.92-1.10) 1.12 (1.03-1.22) 1.13 (1.04-1.23) 1.27 (1.18-1.38) ⬍.001 1.21 (1.13-1.31)
Abbreviations: HR, hazard ratio; NA, not applicable.
aData are given as HR (95% CI) except where indicated otherwise.
bThe multivariate model was adjusted for age (continuous); body mass index (calculated as weight in kilograms divided by height in meters squared) category
(⬍23.0, 23.0-24.9, 25.0-29.9, 30.0-34.9, or ⱖ35.0); alcohol consumption (0, 0.1-4.9, 5.0-29.9, or ⱖ30.0 g/d in men; 0, 0.1-4.9, 5.0-14.9, or ⱖ15.0 g/d in
women); physical activity level (⬍3.0, 3.0-8.9, 9.0-17.9, 18.0-26.9, or ⱖ27.0 hours of metabolic equivalent tasks per week); smoking status (never, past, or
current [1-14, 15-24, or ⱖ25 cigarettes per day]); race (white or nonwhite); menopausal status and hormone use in women (premenopausal, postmenopausal
never users, postmenopausal past users, or postmenopausal current users); family history of diabetes mellitus, myocardial infarction, or cancer; history of
diabetes mellitus, hypertension, or hypercholesterolemia; and intakes of total energy, whole grains, fruits, and vegetables, all in quintiles.
cResults from multivariate model were combined using the random-effects model.
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E5
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

We estimated that 9.3% (95% CI, 5.9%-12.7%) in men
and 7.6% (95% CI, 3.5%-11.7%) in women of total deaths
during follow-up could be prevented if all the partici-
pants consumed fewer than 0.5 servings per day of total
red meat in these cohorts; the estimates were 8.6% (95%
CI, 2.3%-14.7%) in men and 12.2% (95% CI, 3.3%-
21.0%) in women for CVD deaths. However, only 22.8%
of men and 9.6% of women were in the low-risk cat-
egory of total red meat intake.
COMMENT
In these 2 large prospective cohorts of US men and
women, we found that a higher intake of red meat was
associated with a significantly elevated risk of total, CVD,
and cancer mortality, and this association was observed
for unprocessed and processed red meat, with a rela-
tively greater risk for processed red meat. Substitution
of fish, poultry, nuts, legumes, low-fat dairy products,
and whole grains for red meat was associated with a sig-
nificantly lower risk of mortality.
Red meat is a major food source of protein and fat,
and its potential associations with risks of diabetes melli-
tus,1CVD,2cancer,3and mortality4-6 have attracted much
attention. Several studies4,5 have suggested that vegetar-
ians have greater longevity compared with nonvegetar-
ians, but this might not be ascribed to the absence of red
meat only. Sinha et al6showed in the National Institutes
of Health–AARP (formerly known as the American As-
sociation of Retired Persons) study that higher intakes
of red and processed meats were associated with an el-
evated risk of mortality. However, that study did not dis-
tinguish unprocessed and processed red meats and did
not update dietary information during follow-up.
The strengths of the present study include a large sample
size, high rates of long-term follow-up, and detailed and
repeated assessments of diet and lifestyle. All the partici-
pants were health professionals, minimizing potential con-
Table 4. Cancer Mortality According to Red Meat Intake in the Health Professionals Follow-up Study and the Nurses’ Health Study
Variable
Frequency of Consumption Quintilesa
PValue
for Trend
HR (95% CI) for a
1-Serving-per-Day
IncreaseQ1 Q2 Q3 Q4 Q5
Health Professionals Follow-up Study
Total red meat
Cases/person-years, No. 598/152 206 558/153 082 561/152 574 646/153 343 710/152 584 NA NA
Age-adjusted model 1 [Reference] 1.03 (0.91-1.15) 1.05 (0.93-1.18) 1.20 (1.07-1.34) 1.33 (1.20-1.49) ⬍.001 1.17 (1.12-1.22)
Multivariate modelb1 [Reference] 1.05 (0.94-1.18) 1.07 (0.95-1.20) 1.18 (1.05-1.33) 1.24 (1.09-1.40) ⬍.001 1.12 (1.06-1.17)
Unprocessed red meat
Cases/person-years, No. 650/151 745 588/150 121 540/155 255 613/152 008 682/154 661 NA NA
Age-adjusted model 1 [Reference] 1.00 (0.89-1.12) 0.97 (0.86-1.08) 1.06 (0.95-1.18) 1.25 (1.12-1.39) ⬍.001 1.18 (1.11-1.26)
Multivariate modelb1 [Reference] 1.01 (0.90-1.13) 1.03 (0.91-1.15) 1.05 (0.94-1.18) 1.18 (1.05-1.33) ⬍.001 1.13 (1.05-1.21)
Processed red meat
Cases/person-years, No. 669/172 756 487/131 895 580/153 463 589/153 122 748/152 551 NA NA
Age-adjusted model 1 [Reference] 0.97 (0.86-1.09) 1.09 (0.98-1.22) 1.09 (0.97-1.21) 1.28 (1.15-1.42) ⬍.001 1.31 (1.21-1.41)
Multivariate modelb1 [Reference] 1.00 (0.89-1.12) 1.07 (0.96-1.20) 1.07 (0.95-1.20) 1.15 (1.02-1.29) ⬍.001 1.17 (1.07-1.27)
Nurses’ Health Study
Total red meat
Cases/person-years, No. 1264/439 774 1191/443495 1185/440 970 1263/441 727 1488/441 393 NA NA
Age-adjusted model 1 [Reference] 1.04 (0.96-1.13) 1.08 (1.00-1.17) 1.19 (1.10-1.29) 1.39 (1.29-1.50) ⬍.001 1.21 (1.17-1.25)
Multivariate modelb1 [Reference] 1.05 (0.97-1.14) 1.10 (1.01-1.19) 1.15 (1.06-1.25) 1.17 (1.08-1.28) ⬍.001 1.09 (1.05-1.13)
Unprocessed red meat
Cases/person-years, No. 1308/442 572 1222/442671 1120/440 530 1215/432 361 1526/449 225 NA NA
Age-adjusted model 1 [Reference] 1.02 (0.94-1.10) 0.97 (0.90-1.06) 1.09 (1.01-1.18) 1.33 (1.24-1.44) ⬍.001 1.22 (1.17-1.27)
Multivariate modelb1 [Reference] 1.04 (0.96-1.12) 1.03 (0.95-1.12) 1.11 (1.02-1.20) 1.17 (1.08-1.27) ⬍.001 1.09 (1.04-1.14)
Processed red meat
Cases/person-years, No. 1294/444 119 1230/421760 1236/456 687 1204/442 791 1427/442 002 NA NA
Age-adjusted model 1 [Reference] 1.08 (1.00-1.17) 1.11 (1.03-1.20) 1.14 (1.05-1.23) 1.35 (1.25-1.46) ⬍.001 1.41 (1.31-1.52)
Multivariate modelb1 [Reference] 1.05 (0.97-1.14) 1.08 (1.00-1.17) 1.08 (1.00-1.17) 1.14 (1.05-1.23) .001 1.14 (1.05-1.24)
Pooled Resultsc
Total red meat 1 [Reference] 1.05 (0.98-1.12) 1.09 (1.02-1.16) 1.16 (1.08-1.24) 1.19 (1.11-1.28) ⬍.001 1.10 (1.07-1.13)
Unprocessed red meat 1 [Reference] 1.03 (0.97-1.10) 1.03 (0.96-1.10) 1.09 (1.02-1.16) 1.17 (1.10-1.26) ⬍.001 1.10 (1.06-1.14)
Processed red meat 1 [Reference] 1.03 (0.97-1.10) 1.08 (1.01-1.15) 1.08 (1.01-1.15) 1.14 (1.07-1.22) ⬍.001 1.16 (1.09-1.23)
Abbreviations: HR, hazard ratio; NA, not applicable.
aData are given as HR (95% CI) except where indicated otherwise.
bThe multivariate model was adjusted for age (continuous), body mass index (calculated as weight in kilograms divided by height in meters squared) category
(⬍23.0, 23.0-24.9, 25.0-29.9, 30.0-34.9, or ⱖ35.0); alcohol consumption (0, 0.1-4.9, 5.0-29.9, and ⱖ30.0 g/d in men; 0, 0.1-4.9, 5.0-14.9, ⱖ15.0 g/d in
women); physical activity level (⬍3.0, 3.0-8.9, 9.0-17.9, 18.0-26.9, or ⱖ27.0 hours of metabolic equivalent tasks per week); smoking status (never, past, or
current 1-14 cigarettes per day, current 15-24 cigarettes/d, or current ⱖ25 cigarettes/d); race (white or nonwhite); menopausal status and hormone use in women
(premenopausal, postmenopausal never users, postmenopausal past users, or postmenopausal current users); family history of diabetes mellitus, myocardial
infarction, or cancer; history of diabetes mellitus, hypertension, or hypercholesterolemia; and intakes of total energy, whole grains, fruits, and vegetables in all
quintiles.
cResults from the multivariate model were combined using the random-effects model.
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E6
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

founding by educational attainment or differential access
to health care. In addition, the FFQs used in these studies
were validated against multiple diet records.9,10 However,
the measurement errors inherent in dietary assessments were
inevitable, including misclassification of ham or cold cuts
as unprocessed red meat and inaccurate assessment of red
meat content in mixed dishes. Because of the prospective
study design, any measurement errors of meat intake are
independent of study outcome ascertainment and, there-
fore, are likely to attenuate the associations toward the null.16
In the sensitivity analysis accounting for measurement er-
rors, the risk estimates became stronger. Moreover, we cal-
culated cumulative averages for dietary variables to better
represent a person’s long-term diet pattern and to mini-
mize the random measurement error caused by within-
person variation. As expected, the analyses using baseline
diet only yielded weaker associations. We also stopped up-
dating the dietary information after a diagnosis of major
chronic disease assuming that participants could have
changed their diet after receiving the diagnosis. Finally, be-
cause the participants were predominantly non-Hispanic
white health professionals, the generalizability of the ob-
served associations may be limited to similar populations.
Several mechanisms may explain the adverse effect of
red meat intake on mortality risk. Regarding CVD mor-
tality, we previously reported that red meat intake was
associated with an increased risk of coronary heart dis-
ease,2,14 and saturated fat and cholesterol from red meat
may partially explain this association.12 The association
between red meat and CVD mortality was moderately at-
tenuated after further adjustment for saturated fat and
cholesterol, suggesting a mediating role for these nutri-
ents. However, we could not assess whether lean meat
has the same health risks as meat with higher fat con-
tent. Furthermore, dietary iron, particularly heme iron
primarily from red meat, has been positively associated
with myocardial infarction and fatal coronary heart dis-
0 1234
2.5
Hazard Ratio
Total Red Meat Intake, Servings per Day
2.0
1.5
1.0
A
0 1234
2.5
Hazard Ratio
Total Red Meat Intake, Servings per Day
2.0
1.5
1.0
B
Figure 1. Dose-response relationship between red meat intake and risk of
all-cause mortality in the Health Professionals Follow-up Study (A) and the
Nurses’ Health Study (B). The results were adjusted for age (continuous); body
mass index (calculated as weight in kilograms divided by height in meters
squared) category (⬍23.0, 23.0-24.9, 25.0-29.9, 30.0-34.9, or ⱖ35); alcohol
consumption (0, 0.1-4.9, 5.0-29.9, ⱖ30.0 g/d in men; 0, 0.1-4.9, 5.0-14.9, or
ⱖ15.0 g/d in women); physical activity level (⬍3.0, 3.0-8.9, 9.0-17.9,
18.0-26.9, or ⱖ27.0 hours of metabolic equivalent tasks per week); smoking
status (never, past, or current [1-14, 15-24, or ⱖ25 cigarettes per day]); race
(white or nonwhite); menopausal status and hormone use in women
(premenopausal, postmenopausal never users, postmenopausal past users, or
postmenopausal current users); family history of diabetes mellitus, myocardial
infarction, or cancer; history of diabetes mellitus, hypertension, or
hypercholesterolemia; and intakes of total energy, whole grains, fruits, and
vegetables, all in quintiles. Broken lines represent 95% CI.
0.65 0.750.70 0.95
0.850.80 1.00
Hazard Ratios for Total Mortality
0.90
Nuts for unprocessed red meat
Legumes for unprocessed red meat
Low-fat dairy for unprocessed red meat
Whole grains for unprocessed red meat
Poultry for unprocessed red meat
Fish for unprocessed red meat
Nuts for total red meat
Legumes for total red meat
Low-fat dairy for total red meat
Whole grains for total red meat
Poultry for total red meat
Fish for total red meat
Nuts for processed red meat
Legumes for processed red meat
Low-fat dairy for processed red meat
Whole grains for processed red meat
Poultry for processed red meat
Fish for processed red meat
Figure 2. Hazard ratios and 95% CIs (error bars) for total mortality associated with replacement of other food groups for red meat intake. Adjusted for age (continuous);
body mass index (calculated as weight in kilograms divided by height in meters squared) category (⬍23.0, 23.0-24.9, 25.0-29.9, 30.0-34.9, or ⱖ35.0); alcohol
consumption (0, 0.1-4.9, 5.0-29.9, ⱖ30.0 g/d in men; 0, 0.1-4.9, 5.0-14.9, or ⱖ15.0 g/d in women); physical activity level (⬍3.0, 3.0-8.9, 9.0-17.9, 18.0-26.9, or ⱖ27.0
hours of metabolic equivalent tasks per week); smoking status (never, past, or current [1-14, 15-24, or ⱖ25 cigarettes per day]); race (white or nonwhite); menopausal
status and hormone use in women (premenopausal, postmenopausal never users, postmenopausal past users, or postmenopausal current users); family history of
diabetes mellitus, myocardial infarction, or cancer; history of diabetes mellitus, hypertension, or hypercholesterolemia; total energy intake; and the corresponding 2
dietary variables in the models.
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E7
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

ease.17-20 The associations between red meat and CVD mor-
tality were moderately attenuated after additional adjust-
ment for heme iron. This finding suggests that heme iron
intake may partially explain this association, although
some studies using biomarkers of iron status found no
association of ferritin and transferrin saturation levels with
risk of total mortality.21 Unprocessed and processed meats
contain similar amounts of saturated fat and heme iron;
however, other constituents in processed meat, particu-
larly sodium and nitrites, might explain the additional
harm of processed meats. The high sodium content may
increase CVD risk through its effect on blood pres-
sure.22,23 Nitrites and nitrates are frequently used in the
preservation of processed meats, and blood nitrite con-
centrations have been related to endothelial dysfunc-
tion24 and impaired insulin response in adults.25
Regarding cancer mortality, red meat intake has been
associated with increased risks of colorectal cancer and
several other cancers.26 Several compounds in red meat
or created by high-temperature cooking, including N-
nitroso compounds (nitrosamines or nitrosamides) con-
verted from nitrites,27 polycyclic aromatic hydrocar-
bons, and heterocyclic amines,28-30 are potential
carcinogens. Heme iron and iron overload might also be
associated with increased cancer risk through promo-
tion of N-nitroso compound formation,31 increased co-
lonic cytotoxicity and epithelial proliferation,32 in-
creased oxidative stress, and iron-induced hypoxia
signaling.33
In conclusion, we found that greater consumption of
unprocessed and processed red meats is associated with
higher mortality risk. Compared with red meat, other di-
etary components, such as fish, poultry, nuts, legumes,
low-fat dairy products, and whole grains, were associ-
ated with lower risk. These results indicate that replace-
ment of red meat with alternative healthy dietary com-
ponents may lower the mortality risk.
Accepted for Publication: December 20, 2011.
Published Online: March 12, 2012. doi:10.1001
/archinternmed.2011.2287
Correspondence: Frank B. Hu, MD, PhD, Departments
of Nutrition and Epidemiology, Harvard School of Pub-
lic Health, 655 Huntington Ave, Boston, MA 02115 (frank
.hu@channing.harvard.edu).
Author Contributions: Drs Pan and Hu had full access
to all the data in the study and take responsibility for the
integrity of the data and the accuracy of the data analy-
sis. Study concept and design: Pan, Willett, and Hu. Ac-
quisition of data: Manson, Stampfer, Willett, and Hu.
Analysis and interpretation of data: Pan, Sun, Bernstein,
Schulze, Manson, Stampfer, Willett, and Hu. Drafting of
the manuscript: Pan. Critical revision of the manuscript for
important intellectual content: Sun, Bernstein, Schulze,
Manson, Stampfer, Willett, and Hu. Statistical analysis:
Pan, Sun, and Hu. Obtained funding: Manson, Stampfer,
Willett, and Hu. Administrative, technical, and material
support: Manson, Stampfer, Willett, and Hu. Study su-
pervision: Manson, Stampfer, Willett, and Hu.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grants
DK58845, CA55075, CA87969, HL34594, and
1U54CA155626-01 from the National Institutes of Health
and by career development award K99HL098459 from
the National Heart, Lung, and Blood Institute (Dr Sun).
Role of the Sponsors: The funding sources were not in-
volved in the data collection, data analysis, manuscript
writing, and publication.
Online-Only Material: The eTables and eFigure are avail-
able at http://www.archinternmed.com.
Additional Contributions: We are indebted to the par-
ticipants in the HPFS and the NHS for their continuing
outstanding support and to colleagues working in these
studies for their valuable help. In addition, we thank the
following state cancer registries for their help: Alabama,
Arizona, Arkanas, California, Colorado, Connecticut,
Delaware, Florida, Georgia, Idaho, Illinois, Indiana, Iowa,
Kentucky, Louisiana, Maine, Maryland, Massachusetts,
Michigan, Nebraska, New Hampshire, New Jersey, New
York, North Carolina, North Dakota, Ohio, Oklahoma,
Oregon, Pennsylvania, Rhode Island, South Carolina, Ten-
nessee, Texas, Virginia, Washington, and Wyoming.
REFERENCES
1. Pan A, Sun Q, Bernstein AM, et al. Red meat consumption and risk of type 2
diabetes: 3 cohorts of US adults and an updated meta-analysis. Am J Clin Nutr.
2011;94(4):1088-1096.
2. Micha R, Wallace SK, Mozaffarian D. Red and processed meat consumption and
risk of incident coronary heart disease, stroke, and diabetes mellitus: a system-
atic review and meta-analysis. Circulation. 2010;121(21):2271-2283.
3. Zheng W, Lee SA. Well-done meat intake, heterocyclic amine exposure, and can-
cer risk. Nutr Cancer. 2009;61(4):437-446.
4. Fraser GE. Associations between diet and cancer, ischemic heart disease, and
all-cause mortality in non-Hispanic white California Seventh-day Adventists. Am
J Clin Nutr. 1999;70(3)(suppl):532S-538S.
5. Key TJ, Fraser GE, Thorogood M, et al. Mortality in vegetarians and nonvegetar-
ians: detailed findings from a collaborative analysis of 5 prospective studies. Am
J Clin Nutr. 1999;70(3)(suppl):516S-524S.
6. Sinha R, Cross AJ, Graubard BI, Leitzmann MF, Schatzkin A. Meat intake and
mortality: a prospective study of over half a million people. Arch Intern Med. 2009;
169(6):562-571.
7. van Dam RM, Willett WC, Rimm EB, Stampfer MJ, Hu FB. Dietary fat and meat
intake in relation to risk of type 2 diabetes in men. Diabetes Care. 2002;25(3):
417-424.
8. Fung TT, Schulze M, Manson JE, Willett WC, Hu FB. Dietary patterns, meat in-
take, and the risk of type 2 diabetes in women. Arch Intern Med. 2004;164
(20):2235-2240.
9. Hu FB, Rimm E, Smith-Warner SA, et al. Reproducibility and validity of dietary
patterns assessed with a food-frequency questionnaire. Am J Clin Nutr. 1999;
69(2):243-249.
10. Salvini S, Hunter DJ, Sampson L, et al. Food-based validation of a dietary ques-
tionnaire: the effects of week-to-week variation in food consumption. Int J
Epidemiol. 1989;18(4):858-867.
11. Rich-Edwards JW, Corsano KA, Stampfer MJ. Test of the National Death Index
and Equifax Nationwide Death Search. Am J Epidemiol. 1994;140(11):1016-
1019.
12. 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 re-
peated dietary measurements. Am J Epidemiol. 1999;149(6):531-540.
13. Qiu W, Rosner B. Measurement error correction for the cumulative average model
in the survival analysis of nutritional data: application to Nurses’ Health Study.
Lifetime Data Anal. 2010;16(1):136-153.
14. Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major di-
etary protein sources and risk of coronary heart disease in women. Circulation.
2010;122(9):876-883.
15. Spiegelman D, Hertzmark E, Wand HC. Point and interval estimates of partial popu-
lation attributable risks in cohort studies: examples and software. Cancer Causes
Control. 2007;18(5):571-579.
16. Willett WC. Nutritional Epidemiology. 2nd ed. New York, NY: Oxford University
Press; 1998.
17. Ascherio A, Willett WC, Rimm EB, Giovannucci EL, Stampfer MJ. Dietary iron in-
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E8
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from

take and risk of coronary disease among men. Circulation. 1994;89(3):969-974.
18. Klipstein-Grobusch K, Grobbee DE, den Breeijen JH, Boeing H, Hofman A, Wit-
teman JC. Dietary iron and risk of myocardial infarction in the Rotterdam Study.
Am J Epidemiol. 1999;149(5):421-428.
19. van der A DL, Peeters PH, Grobbee DE, Marx JJ, van der Schouw YT. Dietary
haem iron and coronary heart disease in women. Eur Heart J. 2005;26(3):257-
262.
20. Qi L, van Dam RM, Rexrode K, Hu FB. Heme iron from diet as a risk factor for
coronary heart disease in women with type 2 diabetes. Diabetes Care. 2007;
30(1):101-106.
21. Menke A, Muntner P, Ferna´ndez-Real JM, Guallar E. The association of biomark-
ers of iron status with mortality in US adults [published online ahead of print
February 15, 2011]. Nutr Metab Cardiovasc Dis. doi:10.1016/j.numecd.2010.11
.011.
22. Bibbins-Domingo K, Chertow GM, Coxson PG, et al. Projected effect of dietary
salt reductions on future cardiovascular disease. N Engl J Med. 2010;362(7):
590-599.
23. Smith-Spangler CM, Juusola JL, Enns EA, Owens DK, Garber AM. Population
strategies to decrease sodium intake and the burden of cardiovascular disease:
a cost-effectiveness analysis. Ann Intern Med. 2010;152(8):481-487, W170-
W173.
24. Kleinbongard P, Dejam A, Lauer T, et al. Plasma nitrite concentrations reflect the
degree of endothelial dysfunction in humans. Free Radic Biol Med. 2006;40
(2):295-302.
25. Pereira EC, Ferderbar S, Bertolami MC, et al. Biomarkers of oxidative stress and
endothelial dysfunction in glucose intolerance and diabetes mellitus. Clin Biochem.
2008;41(18):1454-1460.
26. World Cancer Research Fund/American Institute for Cancer Research. Food, Nu-
trition, Physical Activity, and the Prevention of Cancer: A Global Perspective. Wash-
ington, DC: American Institute for Cancer Research; 2007.
27. Hughes R, Cross AJ, Pollock JRA, Bingham S. Dose-dependent effect of dietary
meat on endogenous colonic N-nitrosation. Carcinogenesis. 2001;22(1):199-
202.
28. Skog K, Steineck G, Augustsson K, Ja¨gerstad M. Effect of cooking temperature
on the formation of heterocyclic amines in fried meat products and pan residues.
Carcinogenesis. 1995;16(4):861-867.
29. Sinha R, Rothman N, Salmon CP, et al. Heterocyclic amine content in beef cooked
by different methods to varying degrees of doneness and gravy made from meat
drippings. Food Chem Toxicol. 1998;36(4):279-287.
30. Cross AJ, Sinha R. Meat-related mutagens/carcinogens in the etiology of colo-
rectal cancer. Environ Mol Mutagen. 2004;44(1):44-55.
31. Cross AJ, Pollock JR, Bingham SA. Haem, not protein or inorganic iron, is re-
sponsible for endogenous intestinal N-nitrosation arising from red meat. Can-
cer Res. 2003;63(10):2358-2360.
32. Sesink AL, Termont DS, Kleibeuker JH, Van der Meer R. Red meat and colon
cancer: the cytotoxic and hyperproliferative effects of dietary heme. Cancer Res.
1999;59(22):5704-5709.
33. Huang X. Iron overload and its association with cancer risk in humans: evidence
for iron as a carcinogenic metal. Mutat Res. 2003;533(1-2):153-171.
ONLINE FIRST
INVITED COMMENTARY
Holy Cow! What’s Good for You
Is Good for Our Planet
Is red meat bad for you? In a word, yes. In this is-
sue, Pan et al1describe the outcomes from more than
37 000 men from the Harvard Health Professionals
Follow-Up Study and more than 83 000 women from the
Harvard Nurses Health Study who were followed up for
almost 3 million person-years.
This is the first large-scale prospective longitudinal
study showing that consumption of both processed and
unprocessed red meat is associated with an increased risk
of premature mortality from all causes as well as from
cardiovascular disease and cancer. In a related study by
Pan et al,2red meat consumption was also associated with
an increased risk of type 2 diabetes mellitus.
Substitution of red meat with fish, poultry, nuts, le-
gumes, low-fat dairy products, and whole grains was as-
sociated with a significantly lower risk of mortality. We
have a spectrum of choices; it’s not all or nothing.3
Plant-based foods are rich in phytochemicals, biofla-
vonoids, and other substances that are protective. In other
words, what we include in our diet is as important as what
we exclude, so substituting healthier foods for red meat
provides a double benefit to our health.
Pan et al1reported that adjustment for saturated fat,
dietary cholesterol, and heme iron accounted for some
but not all of the risk of eating red meat. Thus, other
mechanisms such as nontraditional risk factors may be
involved.
For example, a recent study by Smith4found that high-
fat, high-protein, low-carbohydrate (HPLC) diets (which
are usually high in red meat, such as the Atkins and Pa-
leolithic diets) may accelerate atherosclerosis through
mechanisms that are unrelated to the classic cardiovas-
cular risk factors. Mice that were fed an HPLC diet had
almost twice the level of arterial plaque as mice that were
fed a Western diet even though the classic risk factors
were not significantly different between groups. The mice
that were fed the HPLC diet had markedly fewer circu-
lating endothelial progenitor cells and higher levels of
nonesterified fatty acids (promoting inflammation) than
mice that were fed the Western diet.5
Therefore, studies of HPLC diets that only examine
their effects on changes in weight, blood pressure, and
lipid levels may not adequately reflect the negative in-
fluence of HPLC diets on health outcomes, such as mor-
bidity and mortality.
There is an emerging consensus among most nutri-
tion experts about what constitutes a healthy way of
eating:
vlittle or no red meat;
vhigh in “good carbs” (including vegetables, fruits,
whole grains, legumes, and soy products in their natu-
ral forms);
vlow in “bad carbs” (simple and refined carbohy-
drates, such as sugar, high-fructose corn syrup, and white
flour);
vhigh in “good fats” (␻-3 fatty acids found in fish oil,
flax oil, and plankton-based oils);
ARCH INTERN MED PUBLISHED ONLINE MARCH 12, 2012 WWW.ARCHINTERNMED.COM
E9
©2012 American Medical Association. All rights reserved.
at Harvard University, on March 12, 2012 www.archinternmed.comDownloaded from