Public Health Nutrition: 14(2), 261–270
Egg consumption and CHD and stroke mortality: a prospective
study of US adults
Carolyn G Scrafford1,2,*, Nga L Tran1,3, Leila M Barraj1and Pamela J Mink4
1Health Sciences Practice, Exponent Inc., 1150 Connecticut Avenue, NW Suite 1100, Washington, DC 20036,
USA:2Department of International Health, Johns Hopkins Bloomberg School of Public Health, Johns Hopkins
University, Baltimore, MD, USA:3Department of Health Policy and Management, Johns Hopkins Bloomberg
School of Public Health, Johns Hopkins University, Baltimore, MD, USA:4Department of Epidemiology, Rollins
School of Public Health, Emory University, Atlanta, GA, USA
Submitted 15 September 2009: Accepted 19 April 2010: First published online 16 July 2010
Objective: To evaluate the relationship between egg consumption and CHD
and stroke mortality using the Third National Health and Nutrition Examination
Survey 1988–1994 (NHANES III) and follow-up survey.
Design: A cross-sectional survey using a stratified, multi-stage probability sample was
analysed, adjusting for survey design. Egg consumption was obtained from the FFQ
and separated into categories of egg intake. Hazard ratios (HR) were calculated for
CHD and stroke mortality using multivariate Cox regression models.
Setting: A health and nutrition survey conducted in the USA from 1988 to 1994 with
follow-up through 31 December 2000.
Subjects: The study population included men and women who were free of CVD and
had completed a FFQ at baseline.
Results: Multivariate models adjusting for health, lifestyle and dietary factors indicated
that ‘high’ egg consumption ($7 times/week v. ,1 time/week) was not associated
with significantly increased CHD mortality (HR51?13, 95% CI 0?61, 2?11 (men);
HR50?92, 95% CI 0?27, 3?11 (women)). There was a statistically significant inverse
association between ‘high’ egg consumption and stroke mortality among men
(HR50?27, 95% CI 0?10, 0?73), but the estimate was imprecise because of sparse
data. We did not observe a statistically significant positive association between ‘high’
egg consumption and CHD or stroke mortality in analyses restricted to individuals
with diabetes, but these analyses may be limited due to the small number of diabetics.
Conclusions: We did not find a significant positive association between egg con-
sumption and increased risk of mortality from CHD or stroke in the US population.
These results corroborate the findings of previous studies.
CHD and stroke are leading causes of death in the USA
and globally. Elevated serum LDL has been identified as a
major risk factor for CHD(1). Data from epidemiological
studies have shown a relationship between dietary cho-
lesterol intake and CHD risk(2,3), and metabolic studies
have shown that an increase in dietary cholesterol resulted
in an increase in plasma total and LDL cholesterol(4,5).
These findings, in part, have led to the guidelines from
the American Heart Association (AHA) recommending
that healthy adults limit their intake of dietary cholesterol
to ,300mg/d. Since a large egg contains about 210mg of
cholesterol, or about 71% of the corresponding recom-
mended daily value, the AHA recommends restricting egg
consumption unless dietary cholesterol intake from other
sources is limited(6).
The AHA guidelines are inconsistent with international
guidelines on egg consumption and cholesterol intake.
The British Heart Foundation recently removed their
advice to limit egg consumption to three per week and
there is currently no restriction on egg consumption(7).
The Food Standards Agency places an emphasis on
reducing saturated fat intake while recommending that
eggs are a good choice in a balanced diet(8). Australia
recommends limiting cholesterol-rich foods but states
that eggs are a good choice if you are healthy and have
normal blood cholesterol levels(9). In addition, country-
specific FAO food-based dietary guidelines indicate that
many countries, including Thailand, Mexico, New Zealand
and Japan, recommend eating eggs regularly as part of
a healthy diet(10).
The AHA rationale is in conflict with a growing body of
scientific literature. Observational epidemiological studies
that used simple regression analyses indicated a positive
relationship between dietary cholesterol and CHD risk,
*Corresponding author: Email email@example.com
r The Authors 2010
whereas results of multiple regression analyses tended to
find no significant positive association(11–15). A cross-
sectional study(16)found that egg consumption was not
associated with elevated serum cholesterol concentra-
tions. In addition, four prospective studies showed that
after adjustment for other potential risk factors, there was
no significant overall association between egg con-
sumption and risk of stroke or CHD(17)or risk of stroke
or CVD(18–20). A review of epidemiological studies by
Kritchevsky and Kritchevsky(21)concluded that there was
no significant positive association between consuming
one or more eggs per day and CHD events, after adjust-
ment for dietary confounders in addition to other known
risk factors. A recent risk apportionment study looking at
the contribution of egg consumption and other modifi-
able lifestyle factors to CHD risk determined that con-
suming one egg per day contributed to ,1% of the CHD
risk in the majority of US adults 25 years of age and above
when adjusting for other risk factors such as smoking,
high BMI, poor dietary patterns and physical inactivity(22).
In contrast, several studies have reported significant
positive associations between egg intake and CHD mortality
among persons with type 2 diabetes(17,19,20,23). In addition,
an inverse association between egg intake and stroke risk
was observed by Sauvaget et al.(24). While there are con-
sistent findings in the scientific literature, several of the
previous prospective studies have not controlled for dietary
factors, in particular saturated fat intake(19,20). Therefore, the
present prospective study to assess the impact of egg con-
sumption on risk of death from CHD and stroke in the US
population with data on potential dietary confounders will
help to fill this gap in the literature. We also evaluated
the relationship between egg consumption and CHD and
stroke in subgroup analyses restricted to men and women
with self-reported type 2 diabetes.
The Third National Health and Nutrition Examination
Survey (NHANES III) is a cross-sectional survey using a
stratified, multi-stage probability sample of civilian, non-
institutionalized individuals 2 months of age and above
residing in the conterminous USA and is described in
detail elsewhere(25). NHANES III was conducted by the
National Center for Health Statistics (NCHS) from 1988 to
1994 and collected nutritional status, lifestyle factors,
health-related behaviours, health status and physical
examination data on 33994 subjects interviewed in their
homes and examined in mobile examination clinics(26).
A semi-quantitative FFQ was administered during the
home interviews among individuals of 17 years of age
and above (n 20050) to assess their qualitative dietary
habits over the past month. Detailed methodology for the
FFQ has been described previously(26).
NCHS conducted a follow-up study with NHANES III
participants linking death records from the National Death
Index (NDI) through 31 December 2000 for individuals
aged 17 years and above at baseline. This linkage was
based on a probabilistic match of the NHANES III records
to the NDI and a sample of death certificates was collected
for review and verification. The matching methodology,
described in detail elsewhere(27), is based on twelve items
that are collected in NHANES III and that are routinely used
by NDI to match, including social security number, first and
last name, birth date, sex, state of birth, race, state of resi-
dence and marital status. Among the eligible population
that could be matched to the NDI (n 20024), there were
3384 confirmed deaths (17%) by 31 December 2000.
For the present study, we included all individuals at
17 years of age and above who had completed the FFQ
at baseline (i.e. 1988–1994) and were eligible for follow-up
(n 20050). Of those eligible for follow-up, there were
twenty-six individuals who did not provide sufficient
information to match to the NDI. Individuals were exclu-
ded from the present analysis if they reported that they had
been told by a doctor that they had congestive heart failure
(n 783), stroke (n 656) or had previously suffered a heart
attack (n 1226), if they did not respond to the relevant egg
consumption question on the FFQ (n 41), or if their total
energy intake was missing or implausibly low (,2510kJ/d
(,600kcal/d)) or high ($20920kJ/d($5000kcal/d); n 3420).
In addition, nine individuals who had zero total person-
years were excluded. After these exclusions, 6833 men
and 8113 women remained eligible for follow-up (Fig. 1).
Definition of mortality endpoints
Definitions for CHD mortality and stroke mortality were
based on the International Classification of Diseases, Tenth
Revision (ICD-10)(28). CHD deaths were defined by ICD-10
codes I20–I25. Stroke deaths were defined by ICD 10 codes
The FFQ included sixty food categories and study parti-
cipants were asked to report their frequency of con-
sumption of each category over the past month. The
question that provided the egg consumption data for
the present analysis was: ‘How many times over the past
month have you consumed eggs including scrambled,
fried, omelettes, hard-boiled and egg salad?’ The imple-
mentation of the FFQ in NHANES III was intended to
collect qualitative dietary data that allows for the assess-
ment of general trends in the participant’s diet. NHANES
III also collected 24h dietary recalls on a subsample of
the study population. However, this represents short-term
intake and would not necessarily be representative of the
participant’s typical diet. For this reason, we decided to
use the FFQ as the method of measuring egg consump-
tion and its association with mortality from CHD and
262CG Scrafford et al.
35.McNamara DJ (2000) Dietary cholesterol and atherosclero-
sis. Biochim Biophys Acta 1529, 310–320.
Fernandez ML (2006) Dietary cholesterol provided by eggs
and plasma lipoproteins in healthy populations. Curr Opin
Clin Nutr Metab Care 9, 8–12.
Herron KL & Fernandez ML (2004) Are the current dietary
guidelines regarding egg consumption appropriate? J Nutr
Lee A & Griffin B (2006) Dietary cholesterol, eggs and
coronary heart disease risk in perspective. Br Nutr Found
Nutr Bull 31, 21–27.
Fielding CJ, Castro GR, Donner C et al. (1986) Distribution
of apolipoprotein E in the plasma of insulin-dependent and
noninsulin-dependent diabetics and its relation to choles-
terol net transport. J Lip Res 27, 1052–1061.
Feskens EJ & Kromhout D (1990) Habitual dietary intake
and glucose tolerance in euglycaemic men: the Zutphen
Study. Int J Epidemiol 19, 953–959.
41.Takashima Y, Iwase Y, Yoshida M et al. (1998) Relationship of
food intake and dietary patterns with blood pressure levels
among middle-aged Japanese men. J Epidemiol 8, 106–115.
Umesawa M, Sato S, Imano H et al. (2009) Relationship of food
intake and dietary patterns with blood pressure levels among
middle-aged Japanese men. Am J Clin Nutr 90, 377–384.
Buzzard M (1998) 24-hour dietary recall and food record
methods. In Nutritional Epidemiology, pp. 50–73 [WC
Willett, editor]. New York: Oxford University Press.
Life Science Research Office (1988) Estimation of Exposure
to Substances in the Food Supply. Report prepared for the
Center for Food Safety and Applied Nutrition, FDA,
Department of Health and Human Services, No. FDA 223-
84-2059. Bethesda, MD: Federation of American Societies
for Experimental Biology.
Agricultural Research Service (2008) National Nutrient Data-
base for Standard Reference, Release 21. Beltsville, MD: ARS;
available at http://www.nal.usda.gov/fnic/foodcomp/search/
270 CG Scrafford et al.