Jiang R, Manson JE, Stampfer MJ, Liu S, Willett WC, Hu FB. Nut and Peanut Butter Consumption and Risk of Type 2 Diabetes in Women. JAMA. 2002;288(20):2554-2560. doi:10.1001/jama.288.20.2554
Author Affiliations: Departments of Nutrition (Drs Jiang, Stampfer, Willett, and Hu) and Epidemiology (Drs Jiang, Manson, Stampfer, Liu, Willett, and Hu), Harvard School of Public Health; the Channing Laboratory (Drs Manson, Stampfer, Willett, and Hu); and the Division of Preventive Medicine (Drs Manson and Liu), Department of Medicine, Harvard Medical School and Brigham and Women's Hospital, Boston, Mass.
Context Nuts are high in unsaturated (polyunsaturated and monounsaturated) fat
and other nutrients that may improve glucose and insulin homeostasis.
Objective To examine prospectively the relationship between nut consumption and
risk of type 2 diabetes.
Design, Setting, and Participants Prospective cohort study of 83 818 women from 11 states in the
Nurses' Health Study. The women were aged 34 to 59 years, had no history of
diabetes, cardiovascular disease, or cancer, completed a validated dietary
questionnaire at baseline in 1980, and were followed up for 16 years.
Main Outcome Measure Incident cases of type 2 diabetes.
Results We documented 3206 new cases of type 2 diabetes. Nut consumption was
inversely associated with risk of type 2 diabetes after adjustment for age,
body mass index (BMI), family history of diabetes, physical activity, smoking,
alcohol use, and total energy intake. The multivariate relative risks (RRs)
across categories of nut consumption (never/almost never, <once/week, 1-4
times/week, and ≥5 times/week) for a 28-g (1 oz) serving size were 1.0,
0.92 (95% confidence interval [CI], 0.85-1.00), 0.84 (0.95% CI, 0.76-0.93),
and 0.73 (95% CI, 0.60-0.89) (P for trend <.001).
Further adjustment for intakes of dietary fats, cereal fiber, and other dietary
factors did not appreciably change the results. The inverse association persisted
within strata defined by levels of BMI, smoking, alcohol use, and other diabetes
risk factors. Consumption of peanut butter was also inversely associated with
type 2 diabetes. The multivariate RR was 0.79 (95% CI, 0.68-0.91; P for trend <.001) in women consuming peanut butter 5 times or more
a week (equivalent to ≥140 g [5 oz] of peanuts/week) compared with those
who never/almost never ate peanut butter.
Conclusions Our findings suggest potential benefits of higher nut and peanut butter
consumption in lowering risk of type 2 diabetes in women. To avoid increasing
caloric intake, regular nut consumption can be recommended as a replacement
for consumption of refined grain products or red or processed meats.
Type 2 diabetes affects approximately 16 million people in the United
States1 and 135 million people worldwide2; the number of people with diabetes will reach an
estimated 300 million worldwide by 2025.2 Because
management of diabetes and its complications such as cardiovascular disease,
amputation, blindness, and renal failure imposes enormous medical and economic
burdens, primary prevention has become a public health imperative.
Recent studies have shown that diet and lifestyle modifications are
important means of preventing type 2 diabetes.3- 5 Evidence
indicates that specific types of dietary fat rather than total fat (as percentage
of energy) intake predict risk of type 2 diabetes.6 Nuts
contain 70% to 80% fat, and most fatty acids in nuts are unsaturated (polyunsaturated
and monounsaturated), which may be beneficial for glucose and insulin homeostasis.
Several studies have shown that a higher intake of monounsaturated and polyunsaturated
fat improves insulin sensitivity.7- 10 A
higher intake of polyunsaturated fat is associated with a lower risk of type
2 diabetes,11 whereas a higher intake of saturated
fat and trans-fat adversely affects glucose metabolism
and insulin resistance8,12- 14 and
thereby may increase the risk of type 2 diabetes.11 Other
components of nuts such as fiber and magnesium decrease insulin demand and
resistance15- 22 and
have been inversely associated with risk of type 2 diabetes.23,24 Nuts
are also a rich source of many vitamins, minerals, and antioxidants and of
plant protein, which could also be beneficial.
Although several components of nuts have been inversely associated with
risk of type 2 diabetes, the overall association of nut consumption with diabetes
risk has not been studied. We therefore examined prospectively the association
between nut consumption and risk of type 2 diabetes in a large cohort of women
from the Nurses' Health Study.
The Nurses' Health Study was established in 1976 when 121 700 female
registered nurses, aged 30 to 55 years and from 11 states, completed a mailed
questionnaire designed to study etiologies of heart disease, cancer, and other
major illnesses.25 Information on lifestyle,
health behaviors, and disease status has been collected on biennially mailed
questionnaires since 1976, and information about vitamin supplement use has
been collected the same way since 1980. Diet was assessed in 1980, 1984, 1986,
1990, and 1994 by using semiquantitative food-frequency questionnaires. For
this analysis, at baseline in 1980, we excluded women with 10 or more food
items left blank or with implausibly high (>3500 kcal/d) or low (<500 kcal/d)
total energy intake. We also excluded women with a history of diabetes, cardiovascular
disease (angina, coronary bypass or angioplasty, myocardial infarction, or
stroke), and cancer (except for nonmelanoma skin cancer). After these exclusions,
83 818 participants remained in the analysis.
The semiquantitative food-frequency questionnaire at baseline (1980)
included 61 foods and was revised and expanded to about twice the number of
foods in subsequent cycles.26,27 Participants
were asked to report their average frequency of consumption of selected foods
and beverages with a specified commonly used unit or portion size during the
previous year. The reproducibility and validity of the dietary questionnaires
are described in detail elsewhere.27
In the 1980 and 1984 dietary questionnaires, we asked the participants
how often, on average, they had consumed nuts (serving size, 28 g [1 oz])
during the previous year: never/almost never, 1 to 3 times a month, once a
week, 2 to 4 times a week, 5 to 6 times a week, once a day, 2 to 3 times a
day, 4 to 6 times a day, or more than 6 times a day. In the 1986, 1990, and
1994 dietary questionnaires, the question for nuts was split into 2 categories:
peanuts and other nuts. Total nut consumption was the sum of the intakes for
peanuts and other nuts. Consumption of peanut butter was assessed in 1980,
1984, 1986, 1990, and 1994, with the same 9 responses as those for nut consumption
(serving size, 15 mL [1 tablespoon]). Although peanuts are botanically classified
as legumes, the fatty acid and nutrient profiles of peanuts are very similar
to other nuts.28 A validation study of the
food-frequency questionnaires in the Nurses' Health Study indicated that nuts
and peanut butter were reported reasonably accurately; the correlation coefficient
was 0.75 between intakes assessed by the 1980 questionnaire and by 4 one-week
diet records for nuts and peanut butter.29 Nutrient
intakes, such as for fats and fiber, were computed by multiplying the consumption
frequency of each food by the nutrient content of the specified portion and
then summing these products across all the food items. The food composition
values were obtained from the Harvard University Food Composition Database
derived from US Department of Agriculture sources30 and
supplemented with manufacturer information.
In 1982, 1988, and 1992, the participants provided information on family
history of diabetes in first-degree relatives. The participants also provided
information on their body weight and cigarette smoking every 2 years during
the follow-up. The correlation coefficient between self-reported weight and
measured weight was 0.96.31 Physical activity
was assessed by a shorter questionnaire in 1980 and 1982.32 More
detailed information on physical activity was first collected in 1986 and
was updated in 1986, 1988, and 1992. We used the cumulative average number
of hours a week spent in moderate or vigorous recreational activities, including
brisk walking, vigorous sports, jogging, bicycling, heavy gardening, and heavy
The outcome was incident type 2 diabetes. To all women who reported
a diagnosis of diabetes on any biennial follow-up questionnaire, we mailed
a supplementary questionnaire regarding symptoms, diagnostic tests, and treatments.
The diagnosis of diabetes was established when at least 1 of the following
criteria was reported on the supplementary questionnaire: (1) 1 or more classic
symptoms (excessive thirst, polyuria, weight loss, hunger, or coma) plus a
fasting plasma glucose concentration of 140 mg/dL (7.8 mmol/L) or higher or
a random plasma glucose concentration of 200 mg/dL (11.1 mmol/L) or higher;
or (2) at least 2 elevated plasma glucose concentrations on different occasions
(fasting, ≥140 mg/dL [7.8 mmol/L]; random, ≥200 mg/dL [11.1 mmol/L];
or random, ≥200 mg/dL [11.1 mmol/L] after at least 2 hours of oral glucose
tolerance testing) in the absence of symptoms; or (3) treatment with hypoglycemic
medication (insulin or oral hypoglycemic agents). The diagnostic criteria
for type 2 diabetes were changed in 1997.33 However,
we used the criteria proposed by the National Diabetes Data Group34 because all our cases were diagnosed before June
1996. We excluded women with type 1 diabetes and women classified as having
gestational diabetes only. A validation study in a subsample of this cohort
demonstrated that our supplementary questionnaire is highly reliable in confirming
diabetes diagnoses.35 Among a random sample
of 84 women classified by our criteria as having type 2 diabetes according
to the information reported on the supplementary questionnaire, medical records
were available for 62. An endocrinologist blinded to the information reported
on the questionnaire reviewed the records. The diagnosis of type 2 diabetes
was confirmed in 61 (98%) of the 62 women.
Each participant contributed follow-up time from the date of returning
the 1980 questionnaire to the date of first diagnosis of type 2 diabetes,
death, or June 1, 1996. Women were excluded from subsequent follow-up if they
developed diabetes. In the primary analyses, incidence of type 2 diabetes
was related to nut consumption at baseline. In further analyses, incidence
of type 2 diabetes was related to the cumulative average of nut consumption
from dietary questionnaires administered in 1980, 1984, 1986, 1990, and 1994.36 We separated women into 4 categories according to
their frequency of nut consumption: never/almost never, less than once a week
(1-3 times/month), 1 to 4 times a week, and at least 5 times a week. Relative
risks (RRs) of diabetes were estimated from Cox proportional hazards models.37 Potential confounding variables including body mass
index (BMI), family history of diabetes, physical activity, smoking, alcohol
consumption, and multivitamin supplements were updated during follow-up. In
multivariate models, we adjusted for age, BMI, family history of diabetes,
physical activity, smoking, alcohol use, and total energy intake. We also
adjusted for several dietary confounding variables such as glycemic load,
multivitamin use, and intakes of polyunsaturated fat, saturated fat, trans-fat, cereal fiber, magnesium, whole grains, vegetables,
fruits, and fish.
In a secondary analysis, we performed a propensity analysis38 in which we used a logistic regression model to predict
nut consumption (≥5 times/week vs never/almost never) from a multitude
of dietary and lifestyle factors. We then examined the association between
nut consumption and diabetes risk from a Cox proportional hazards model adjusting
for the predicted propensity scores.
We also calculated mean weight change from 1980 to 1996 for women according
to frequency of nut consumption at baseline (4 categories) by using general
linear models (least square means).39 In this
analysis, we adjusted for age, family history of diabetes, physical activity,
smoking, alcohol use, and baseline weight and excluded women who developed
cancer, heart disease, or diabetes during the follow-up.
All P values were 2-sided. Tests for trend
were conducted using the median value for each category of nut consumption
analyzed as a continuous variable in the regression models. All analyses were
performed with SAS version 6.12 software (SAS Institute, Cary, NC).
At baseline in 1980, about 35% of women in this cohort reported consuming
nuts almost never; 36%, consuming them less than once a week; 24%, 1 to 4
times a week; and 5%, at least 5 times a week. Women who consumed more nuts
generally weighed less (Table 1).
Women with frequent nut consumption were less likely to smoke and more likely
to exercise. Nut consumption was positively associated with intakes of polyunsaturated
fat, dietary fiber, magnesium, alcohol, and multivitamin supplements and inversely
associated with glycemic load and intake of trans-fat.
Consumption of vegetables and fruits was similar for women with frequent nut
consumption and those who rarely ate nuts, but women who consumed more nuts
generally ate less meat and refined grain products.
We documented 3206 incident cases of type 2 diabetes during 1 282 892
person-years of follow-up from 1980 to 1996. The nut consumption at baseline
was used to classify women into 4 categories (never/almost never, <once/week,
1-4 times/week, and ≥5 times/week) so that the age-adjusted RR of diabetes
was 0.55 (95% confidence interval [CI], 0.45-0.66), comparing women who ate
nuts at least 5 times a week with those who never/almost never ate nuts (P for trend <.001) (Table 2). In multivariate models, BMI was the strongest confounder.
The RR was attenuated to 0.74 (95% CI, 0.61-0.89; P for
trend <.001) after BMI was added to the model by using it as a categorical
variable (8 categories) and was 0.72 (95% CI, 0.59-0.87; P for trend <.001) when BMI was used as a continuous variable. The
RR was virtually unchanged after further control for family history of diabetes,
physical activity, smoking, alcohol consumption, and total energy intake.
After controlling for other dietary variables such as glycemic load, multivitamin
use, and intakes of polyunsaturated fat, saturated fat, trans-fat, cereal fiber, magnesium, whole grains, vegetables, fruits,
and fish, the RR did not appreciably change (0.71; 95% CI, 0.57-0.87). Updated
analyses using the cumulative average of nut consumption yielded similar results
(multivariate RR, 0.76; 95% CI, 0.59-0.97; P for
trend = .001).
In the secondary analysis controlling for propensity scores, those who
ate nuts at least 5 times a week still had a lower diabetes risk compared
with those who never/almost never ate nuts (RR, 0.73; 95% CI, 0.56-0.96).
To examine further whether the relationship between nut consumption
and type 2 diabetes risk was independent of other potential risk factors for
type 2 diabetes, we conducted multivariate analyses within strata defined
by levels of these factors. We found no apparent modification of the relationship
by these factors, and the inverse association persisted in all subgroups (Table 3).
We also examined the relationship between consumption of peanut butter
and risk of type 2 diabetes. Frequent consumption of peanut butter was associated
with a significantly reduced risk of type 2 diabetes (Table 4). The multivariate RR was 0.79 (95% CI, 0.68-0.91), comparing
women who ate peanut butter at least 5 times a week with those who never/almost
never ate peanut butter.
To address the concern that higher nut consumption may lead to more
weight gain, we calculated average weight change during 16 years of follow-up
according to frequency of nut consumption at baseline. After adjustment for
age, family history of diabetes, physical activity, smoking, alcohol use,
and baseline weight, the average weight gain across categories of nut consumption
(never/almost never, <once/week, 1-4 times/week, ≥5 times/week) was
not significantly different (6.5, 6.4, 6.4, and 6.3 kg, respectively).
In this large prospective cohort study of women, we found that consumption
of nuts and peanut butter was inversely associated with risk of type 2 diabetes,
independent of known risk factors for type 2 diabetes, including age, obesity,
family history of diabetes, physical activity, smoking, and dietary factors.
The inverse association with nuts persisted in all subgroup analyses.
The major concern of our analysis is residual confounding by body weight
because obesity is the most important determinant of type 2 diabetes. In our
analyses, we adjusted for BMI by using detailed categories and continuous
variables, and the results did not change appreciably. Although we cannot
rule out the possibility of residual confounding by other potential risk factors,
it is unlikely that they can explain the observed inverse association. The
likelihood of bias is minimized because of the prospective study design, high
follow-up rate, and repeated measures of diet with validated food-frequency
questionnaires. Another issue deserving attention is that the diagnostic criteria
for type 2 diabetes used in this study were changed in 1997 so that lower
fasting glucose levels (≥126 mg/dL [7.0 mmol/L]) would now be considered
the diagnostic cut point.33 If the new criteria
were used, some women classified as being without diabetes would have been
reclassified as having diabetes. However, this change would not explain our
results, because inclusion of those with diabetes in the group without diabetes
would tend to weaken the association.
Although the relationship between nut consumption and risk of type 2
diabetes has not been evaluated previously, several studies have examined
the relationship between the major constituents of nuts (eg, specific types
of fat) and insulin resistance and type 2 diabetes.6- 14,24,40- 52 Most
clinical and epidemiological studies have observed no effect of total fat
intake on insulin sensitivity and risk of type 2 diabetes.11,24,42,43,47- 49
The findings on types of dietary fat have been
inconsistent.8,11,12,24,40,42- 44,50- 52 Some
of the inconsistency could be due to methodological limitations in many studies,
such as small sample size, short duration, crude dietary assessment methods
without documented validity, and absence of control for confounding (eg, other
types of fats, fiber intake, obesity, and physical activity). More recent
clinical and epidemiological studies with better design suggest that specific
types of fat rather than total fat as percentage of energy play an important
role in the development of type 2 diabetes.6,8,11 A
multicenter study involving a 3-month intervention among 162 healthy men and
women showed that a diet high in saturated fat (18% of energy) decreased insulin
sensitivity compared with a diet high in monounsaturated fat (21% of energy)
with the same total fat content.8 An analysis11 from the Nurses' Health Study with 14 years of follow-up
showed that women in the highest quintile of vegetable fat intake had a 40%
lower risk of type 2 diabetes than those in the lowest quintile. The RR comparing
extreme quintiles of polyunsaturated fat intake was 0.75 (95% CI, 0.65-0.88).
An Italian study showed that a higher consumption of oils consisting mostly
of polyunsaturated fat was associated with lower fasting plasma concentrations
The mechanisms by which specific types of dietary fat affect insulin
sensitivity are not well understood. It has been shown that the fatty acid
composition of the phospholipids in the skeletal muscle cell membranes is
directly related to insulin sensitivity in humans.53 A
specific fatty acid in cell membranes could influence insulin action through
altering insulin receptor binding or affinity and influencing ion permeability
and cell signaling. Changes in dietary fatty acid composition alter fatty
acid composition of the phospholipids in cell membranes, perhaps modulating
insulin action and sensitivity.
Nuts are also rich in fiber and magnesium and have a relatively low
glycemic index. In several clinical studies, high-fiber diets decreased insulin
demand among patients with type 2 diabetes.15- 17 Also,
metabolic studies suggest an inverse association between intracellular magnesium
and insulin resistance,18,22 and
magnesium supplementation increased insulin sensitivity among patients with
type 2 diabetes19,20 and among
healthy subjects.21 Higher intakes of fiber
and magnesium and foods with a low glycemic index have been associated with
reduced risk of type 2 diabetes in several prospective studies.23,24,54,55
Nuts may protect against type 2 diabetes through additional mechanisms.
The persistence of an association when intakes of types of fats, fiber, and
magnesium were in the model indicates that the apparent benefit of nuts was
not explained entirely by content of fats, fiber, and magnesium. Thus, other
constituents of nuts such as vitamins, minerals, antioxidants, and plant protein
or interactions among these factors may also play important roles in reducing
risk of type 2 diabetes.
There have been concerns that frequent nut consumption may result in
weight gain and increased risk of coronary heart disease because of the high
fat content. However, in our cohort, we did not find an appreciable association
between nut consumption and weight change. Also, several large prospective
studies56- 59 have
consistently found an inverse association between nut consumption and the
risk of coronary heart disease. The epidemiological findings are supported
by several clinical studies28,60- 63 in
which diets high in nuts had beneficial effects on blood lipids. These results
contradict the conventional wisdom that intake of high-fat foods leads to
obesity and heart disease. Given the observed inverse association between
nuts and risk of coronary heart disease as well as type 2 diabetes, it is
advisable to recommend regular nut consumption as a replacement for refined
grain products64 or red or processed meats,65 which would avoid increasing caloric intake.
In conclusion, higher consumption of nuts and peanut butter was associated
with a lower risk of type 2 diabetes in this large cohort study of women.
Our data, combined with other clinical and epidemiological data, support potential
benefits of increasing nut consumption in reducing type 2 diabetes risk.