American Journal of Epidemiology Vol. 179, No.

11
© The Author 2014. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of DOI: 10.1093/aje/kwu052
Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. Advance Access publication:
April 30, 2014

Original Contribution

Overweight in Early Adulthood, Adult Weight Change, and Risk of Type 2 Diabetes,
Cardiovascular Diseases, and Certain Cancers in Men: a Cohort Study

Renée de Mutsert*, Qi Sun, Walter C. Willett, Frank B. Hu, and Rob M. van Dam
* Correspondence to Dr. Renée de Mutsert, Department of Clinical Epidemiology, C7-98, Leiden University and Medical Center, P.O. Box 9600,
2300 RC Leiden, The Netherlands (e-mail: r.de_mutsert@lumc.nl).

Initially submitted August 6, 2013; accepted for publication February 25, 2014.

The relative importance of overweight after childhood and excess weight gain during adulthood remains unclear.
In 39,909 male participants of the Health Professionals Follow-Up Study who were 40–75 years of age in 1986 and
were followed until 2008, we documented 8,755 incident cases of obesity-related chronic diseases (type 2 diabetes
mellitus, cardiovascular diseases, and colorectal, renal, pancreatic, and esophageal cancers). We calculated
composite and cause-specific hazard ratios using a model that included body mass index (BMI; weight (kg)/height
(m)2) at 21 years of age, weight change since age 21 years, smoking, alcohol consumption, and family histories of
myocardial infarction, colon cancer, and diabetes. Compared with a BMI at 21 years of 18.5–22.9, the composite
hazard ratio for a BMI of 23–24.9 was 1.22 (95% confidence interval (CI): 1.16, 1.29), that for a BMI of 25.0–27.4
was 1.57 (95% CI: 1.48, 1.67), that for a BMI of 27.5–29.9 was 2.40 (95% CI: 2.17, 2.65), and that for a BMI ≥30.0
was 3.15 (95% CI: 2.76, 3.60). The composite hazard ratios for adult weight gain compared with a stable weight
were 1.12 (95% CI: 1.03, 1.22) for a gain of 2.5–4.9 kg, 1.41 (95% CI: 1.31, 1.52) for a gain of 5–9.9 kg, 1.72
(95% CI: 1.59, 1.86) for a gain of 10–14.9 kg, and 2.45 (95% CI: 2.27, 2.63) for a gain ≥15 kg. Adiposity in early
adulthood and adult weight gain were both associated with marked increases in the risk of major chronic diseases in
middle-aged and older men, and these associations were already apparent at modest levels of overweight and
weight gain.

cardiovascular disease; diabetes mellitus; epidemiology; mortality; obesity

Abbreviations: BMI, body mass index; CI, confidence interval; CVD, cardiovascular disease.

The high and increasing prevalence of overweight and obe-
sity in children and adolescents (1) is a major public health
concern. Obesity in adolescence and early adulthood has
been related to premature morbidity and mortality in adulthood
(2–11). The health consequences of moderate overweight in
early adulthood are less clear. Furthermore, most studies on
obesity in adolescence have focused on all-cause mortality
(2, 7–10) or a single disease (5, 12–14) as the outcome; few
studies have evaluated the impact of excess adiposity in early
adulthood on the incidence of a variety of chronic diseases
while taking into account competing events (3, 15).
To interpret the findings about adiposity in early adulthood
and the risk of chronic diseases at older ages, it is important
to consider how adult weight changes may modify this
association. Data on the association between weight gain
since adolescence and health outcomes in middle or older
age are limited (9, 13, 16–21). Particularly, the relative im-
portance of adiposity in early adulthood and adult weight
gain for the onset of chronic diseases in middle and older
age remains unclear.
We therefore prospectively evaluated the associations of
adiposity in early adulthood and subsequent weight change
with the incidence of chronic diseases in middle-aged and
older men who were followed for 22 years. We studied major
chronic diseases that have been consistently associated with
adult obesity so that we could clearly distinguish the associ-
ations of moderate overweight and weight gain in early adult-
hood with those diseases.

1353 Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 1354 de Mutsert et al.

Table 1. Age-Standardized Characteristics of Participants (n = 39,909) by Categories of BMI at 21 Years of Age, Health Professionals Follow-up
Study, United States, 1986a

BMI at 21 Years of Agec

Characteristic and Age <18.5 18.5–22.9 23.0–24.9 25.0–27.4 27.5–29.9 ≥30.0
at Baseline, yearsb (n = 1,296) (n = 19,065) (n = 10,669) (n = 6,674) (n = 1,491) (n = 714)
% Mean (SD) % Mean (SD) % Mean (SD) % Mean (SD) % Mean (SD) % Mean (SD)

Age at baseline (1986), 57.1 (10.1) 54.8 (9.9) 52.7 (9.2) 52.3 (9.1) 52.1 (9.0) 52.1 (9.1)
years
BMI at baseline (1986) 22.8 (2.7) 24.2 (2.4) 25.8 (2.3) 27.5 (3.0) 29.3 (3.5) 33.6 (8.9)
<50.0 22.0 (2.6) 23.7 (2.1) 25.6 (2.2) 27.3 (2.9) 29.4 (3.6) 32.8 (7.6)
50.0–64.9 23.3 (2.7) 24.5 (2.4) 26.0 (2.4) 27.8 (3.0) 29.3 (3.5) 33.5 (8.4)
≥65.0 23.4 (2.8) 24.6 (2.5) 26.0 (2.5) 27.1 (2.9) 28.9 (3.3) 36.2 (12.9)
BMI at 21 years of age 17.6 (0.7) 21.1 (1.2) 23.8 (0.5) 25.9 (0.7) 28.5 (0.7) 33.6 (6.2)
<50.0 17.6 (0.7) 21.2 (1.1) 23.8 (0.5) 25.9 (0.7) 28.5 (0.7) 32.9 (5.0)
50.0–64.9 17.6 (0.7) 21.1 (1.2) 23.8 (0.5) 25.9 (0.7) 28.5 (0.7) 33.3 (5.9)
≥65.0 17.5 (0.7) 20.9 (1.2) 23.8 (0.5) 25.9 (0.7) 28.6 (0.6) 36.4 (8.8)
Weight change since 16.9 (8.9) 9.6 (7.7) 6.5 (7.4) 5.2 (9.2) 2.5 (11.1) −0.9 (16.2)
21 years of age, kg
<50.0 14.4 (8.1) 7.8 (6.8) 5.8 (7.0) 4.7 (9.0) 2.8 (11.3) −1.0 (15.1)
50.0–64.9 18.4 (9.2) 10.7 (7.9) 7.0 (7.6) 6.2 (9.3) 2.6 (11.1) −0.1 (17.5)
≥65.0 18.6 (8.7) 11.3 (8.2) 6.9 (7.7) 3.9 (9.0) 1.1 (10.4) −3.4 (15.1)
Family history of MI 31 32 33 33 35 39
<50.0 28 29 29 31 32 35
50.0–64.9 33 36 36 35 36 43
≥65.0 33 31 33 35 38 35
Family history of colon 9 8 8 9 9 8
cancer
<50.0 8 6 7 8 6 8
50.0–64.9 10 9 9 10 11 9
≥65.0 10 8 10 11 10 10
Family history of 13 14 14 14 16 16
diabetes
<50.0 12 12 12 12 13 16
50.0–64.9 14 14 15 15 18 19
≥65.0 15 16 15 18 20 10
Table continues

METHODS
Study design and population
The Health Professional Follow-up Study is an ongoing
prospective cohort study of 51,529 male dentists, optome-
trists, osteopaths, pharmacists, podiatrists, and veterinarians
aged 40–75 years at baseline in 1986. The Human Subjects
Committee of the Harvard School of Public Health approved
this study, and all participants gave written consent.
At baseline, the men completed a questionnaire that col-
lected information about their medical history, diet, and life-
style factors. Information on medical history was updated via
mailed questionnaires every second year. For the present
analysis, we excluded 8,385 men who reported a history of
one of the studied end points or any chronic disease at base-
line that may be associated with either obesity or weight loss
(2,148 with cancer, 1,605 with diabetes, 2,185 with cardiovas-
cular disease (CVD), 2,403 with asthma or chronic obstructive
pulmonary diseases, and 44 with renal failure). Furthermore,
5 men for whom we were missing date of death, 15 men for
whom we were missing date of birth, and 3 men who were out-
side the baseline age range of 40–75 years were consecutively
excluded. Of the remaining participants, 1,888 men were miss-
ing information on the initial questionnaires about recalled
weight at 21 years of age and 1,283 men were missing infor-
mation on weight change during the 5 years before baseline.
These participants were excluded, as were 32 men who had
a body mass index (BMI, defined as the weight in kilograms
divided by the square of the height in meters) below 15 at 21
years of age and 9 men who had a BMI below 15 at the base-
line of the study. After these exclusions, 39,909 men were
included in the analysis.

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 Overweight, Weight Change, and Chronic Disease Risk 1355

Table 1. Continued

BMI at 21 Years of Agec

Characteristic and Age <18.5 18.5–22.9 23.0–24.9 25.0–27.4 27.5–29.9 ≥30.0
at Baseline, yearsb (n = 1,296) (n = 19,065) (n = 10,669) (n = 6,674) (n = 1,491) (n = 714)
% Mean (SD) % Mean (SD) % Mean (SD) % Mean (SD) % Mean (SD) % Mean (SD)

Current smoking 12 10 10 10 11 11
<50.0 11 9 9 11 13 14
50.0–64.9 13 11 11 10 10 8
≥65.0 10 8 8 9 8 14
Past smoking 41 42 41 42 46 47
<50.0 31 33 36 37 38 40
50.0–64.9 47 47 44 45 51 54
≥65.0 51 51 49 49 51 48
Physical activity, MET 15.2 (21.3) 19.1 (24.1) 21.1 (25.8) 20.9 (27.5) 19.3 (28.1) 17.8 (22.7)
hours/week
<50.0 16.1 (24.2) 20.8 (25.9) 22.8 (28.3) 23.1 (30.1) 20.5 (28.5) 21.2 (26.0)
50.0–64.9 14.7 (19.3) 18.1 (22.8) 19.9 (23.2) 19.7 (24.0) 18.3 (25.9) 16.1 (21.1)
≥65.0 14.4 (18.9) 17.8 (22.7) 19.9 (26.1) 18.7 (29.3) 18.8 (33.0) 13.9 (15.2)
Alcohol consumption, 11.7 (17.0) 11.7 (15.4) 11.5 (15.4) 11.2 (15.3) 10.1 (14.2) 9.9 (14.8)
g/day
<50.0 9.8 (15.0) 10.4 (14.1) 10.6 (14.0) 10.4 (14.6) 9.4 (13.1) 8.4 (12.3)
50.0–64.9 13.0 (17.7) 12.7 (16.2) 12.2 (16.3) 11.8 (15.9) 11.1 (15.0) 11.3 (16.9)
≥65.0 13.5 (19.7) 11.9 (16.2) 12.1 (16.6) 11.3 (15.4) 9.0 (14.7) 10.1 (14.3)

Abbreviations: BMI, body mass index; MET h/wk, metabolic equivalent hours per week; MI, myocardial infarction; SD, standard deviation.
a
P < 0.001 for all continuous variables across categories of BMI (weight (kg)/height (m)2) at 21 years of age, except for age at baseline in
participants younger than 50 years of age (P = 0.514) and for MET h/wk in participants 65 years of age or older (P = 0.002) using analysis of
variance F test. P < 0.001 for all categorical variables across categories of BMI at age 21 years, except for a family history of colon cancer
(overall P = 0.086; for participants 50.0–64.9 years of age, P = 0.023), family history of diabetes (overall P = 0.140; for participants <50.0 years
of age, P = 0.042), current smoking (overall P = 0.168; for participants 50.0–64.9 years of age, P = 0.002), and past smoking (overall P = 0.002;
for participants ≥65.0 years of age, P = 0.005), using χ2 test.
b
Characteristics were reported as current characteristics on the enrollment questionnaire in 1986 when participants were 40–75 years of age
except for BMI at age 21 years, which was calculated with the height reported in 1986 and the recalled weight at 21 years of age.
c
Data (except for age) have been standardized to the age distribution of the study population. The age categories are <50.0, 50.0–64.9, and ≥65 years.

Data collection
At study baseline, the participants reported their current
height and weight, their recalled weight at 21 years of age,
and the weight change that occurred in the 5 years before the
study baseline. In a validation study among 123 participants,
the Pearson correlation coefficient between self-reported
weight and the mean of 2 standardized technician-measured
weights was 0.97, and the mean self-reported weight was
1.06 kg (95% confidence interval (CI): 0.44, 1.71) lower
than the measured weight (22). BMI at the age of 21 years
was calculated as the weight at 21 years in kilograms divided
by the square of the height reported in 1986 in meters. We
defined adult weight change as the change in weight between
the age of 21 years and the baseline of the study in 1986.
The baseline questionnaire also included questions about
whether the participants had first-degree family members
with a history of diabetes, myocardial infarction, or colon
cancer and about lifestyle factors, such as cigarette smoking
and alcohol consumption. Participants were asked about their
mean weekly time spent engaged in 8 recreational activities
during the previous year (23). The total energy expended
during physical activity was expressed in hours per week of
metabolic equivalents. Participants reported their average
consumption of approximately 130 foods and beverages dur-
ing the previous year on a validated semiquantitative food
frequency questionnaire (24).
Obesity-related diseases
As the primary end point of this study, we defined a com-
posite outcome of the first occurrence of fatal or nonfatal
chronic obesity-related disease. In this outcome variable,
we included major chronic diseases that have been consis-
tently associated with obesity: CVD (defined as myocardial
infarction or stroke) (25), type 2 diabetes mellitus (25, 26),
and certain cancers (cancers of the rectum, colon, kidney,
pancreas, and esophagus) (25, 26), as well as death from
these diseases.
Every 2 years after the baseline of the study, participants
were asked to report new diseases that had been diagnosed
during the past 2 years on their follow-up questionnaires. Re-
sponse rates for these questionnaires have been consistently

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 1356 de Mutsert et al.
over 90%. We asked all men who reported incident diagnoses
of CVD or cancer to confirm the report and provide permis-
sion to review their medical records. Study investigators who
were unaware of the subjects’ risk factor status reviewed the
medical records to confirm the diagnoses. All men who re-
ported a diagnosis of diabetes on any of the biennial follow-
up questionnaires received a supplementary questionnaire
about symptoms, diagnostic tests, and medication. Deaths dur-
ing follow-up were reported by family members, coworkers, or
postal authorities or were identified through the National Death
Index (27, 28). Physicians reviewed death certificates and
hospital or pathology reports to classify individual causes
of death and were unaware of participants’ reported question-
naire results. A detailed description of definitions and confir-
mation of the diagnoses can be found in Web Appendix 1,
available at http://aje.oxfordjournals.org/.
Statistical analysis
BMI at 21 years of age was categorized according to the
World Health Organization criteria, with 2 additional cutoff
points at 23 and 27.5 as recommended by a World Health Or-
ganization expert consultation (29). This allowed a detailed
examination of the associations of BMI with the outcome
parameters within the “normal weight” and “overweight”
ranges (<18.5, 18.5–22.9 (reference category), 23.0–24.9,
25.0–27.4, 27.5–29.9, and ≥ 30.0). Adult weight change
was calculated as the difference between the reported weight
at the baseline of the study in 1986 and the recalled weight at
the age of 21 years and was grouped into 7 categories, using
stable weight (<2.5-kg weight change) as the reference cate-
gory (other weight-change categories: <−5.0 kg (n = 1,779);
−5.0 to −2.5 kg (n = 1,406); −2.4 to 2.4 kg (n = 7,298); 2.5 to
4.9 kg (n = 5,909); 5.0 to 9.9 kg (n = 9,823); 10.0 to 14.9 kg
(n = 6,546); and ≥15.0 kg (n = 7,148)). We calculated
age-standardized mean values and proportions of characteris-
tics of the participants at the baseline of the study for catego-
ries of BMI at age 21 years and tested differences in baseline
characteristics between BMI categories with 1-way analysis
of variance for continuous variables and with χ2 tests for cat-
egorical variables. Because participants could have suffered
from multiple diseases, which could possibly result in compet-
ing risks, we studied a composite end point of the first occur-
rence of fatal or nonfatal diabetes, CVD, or obesity-related
cancer. In addition, we studied the separate outcomes using
cause-specific hazards, censoring on any prior event that was
not the event of interest (30). In these analyses of the cause-
specific hazards, only persons without any prior event were
included in the separate outcome variables. For example, the
participants who first experienced a diagnosis of diabetes
and later during follow-up experienced a cardiovascular event
were counted as cases in the cause-specific hazard of diabetes
but were censored in the cause-specific hazard of CVDs.
Time of follow-up was defined as the number of months
between the month of returning the baseline questionnaire
in 1986 and the month during which the first major chronic
obesity-related diseases was reported, the month of death, or
the end of the follow-up on January 31, 2008, whichever oc-
curred first. We calculated incidence rates with 95% confi-
dence intervals for both the composite end point and the
Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075
by Ldaseptivany@gmail.com
on 02 December 2017
cause-specific outcomes. We performed Cox proportional
hazards regression using age in months as the time scale
and the month of baseline questionnaire return as the entry
time to estimate the composite and cause-specific hazard ra-
tios and 95% confidence intervals for categories of BMI at
age 21 years and adult weight change as compared with the
reference categories. We evaluated the proportionality of the
hazards during time (i.e., age) by including an interaction
term for the exposure variables with age and tested with
the log likelihood ratio statistic whether inclusion of these in-
teraction terms improved the fit of the models.
We evaluated potential confounding by cigarette smoking,
alcohol consumption, physical activity level, family history
of myocardial infarction, colon cancer and diabetes, and die-
tary factors at baseline (quintiles of energy-adjusted intake of
cereal fiber, vegetables, fruit, ω-3 fatty acids, red meat, trans
fat, and ratio of polyunsatured fat to saturated fat). In separate
analyses, we mutually adjusted for BMI at age 21 years and
adult weight change to examine their independent contributions.
All analyses were also performed for all-cause mortality as
the outcome.
To test for trends, we calculated the median values of BMI
at 21 years of age within each category of BMI at 21 years and
modeled these median values as a continuous variable in all
models with BMI at 21 years. Similarly, we modeled the me-
dians of weight change within each adult weight-change cat-
egory as a continuous variable in all adult weight-change
models. To minimize confounding by cigarette smoking,
we also conducted the analyses in never smokers. To reduce
the impact of potentially unintentional weight loss, we also
conducted the analyses after excluding the weight change
that occurred during the 5 years before baseline. In addition,
we performed the analyses of weight change in 3 strata of
BMI at 21 years of age (<20.0, 20.0–24.9, and ≥25.0). Fi-
nally, we examined joint associations of BMI at 21 years of
age in 3 categories (18.5–22.9, 23–24.9, and ≥25; partici-
pants with a BMI <18.5 were excluded) and subsequent
weight change in 4 categories (<−2.5 kg, −2.5 to 2.4 kg,
2.5 to 9.9 kg, and ≥10.0 kg). We used SAS software, version
9 (SAS Institute, Inc., Cary, North Carolina) for all analyses.
RESULTS
Participants had a mean age of 53.8 (standard deviation,
9.6) years, a mean BMI at age 21 years of 23.0 (standard de-
viation, 2.9), and a mean subsequent weight change of 7.9
(standard deviation, 8.8) kg. The baseline characteristics of
the study population (in 1986) according to BMI at age 21
years are shown in Table 1. Men who had a low BMI at 21
years of age were older and less physically active at baseline,
whereas men who were overweight or obese at 21 years of
age were more likely to be former smokers and to have a fam-
ily history of myocardial infarction or diabetes. The mean
adult weight gain was larger in men who had a lower BMI
at 21 years of age. Nevertheless, overweight and obesity
tended to track over time, with participants who had higher
BMIs at baseline being in the highest categories of BMI at
21 years of age. Compared with participants who were youn-
ger at baseline, participants who were 50–65 years of age at
Am J Epidemiol. 2014;179(11):1353–1365
 Overweight, Weight Change, and Chronic Disease Risk 1357

Table 2. Hazard Ratios of Fatal and Nonfatal Obesity-Related Diseases by Categories of BMIa at 21 Years of Age (n = 39,909), Health
Professionals Follow-up Study, United States, 1986–2008

HR Adjusted for
No. of Age-Adjusted Multivariate
Disease and BMI Category 95% CI 95% CI Subsequent 95% CI
Events HR HRb
Weight Changec

Obesity-related diseasesd,e
<18.5 353 1.29 1.16, 1.44 1.28 1.15, 1.43 0.99 0.88, 1.10
18.5–22.9 3,969 1.00 Referent 1.00 Referent 1.00 Referent
23.0–24.9 2, 145 1.05 0.99, 1.10 1.05 1.00, 1.11 1.22 1.16, 1.29
25.0–27.4 1,571 1.29 1.22, 1.37 1.29 1.21, 1.37 1.57 1.48, 1.67
27.5–29.9 460 1.86 1.69, 2.05 1.81 1.64, 1.99 2.40 2.17, 2.65
≥30.0 257 2.34 2.06, 2.65 2.27 2.00, 2.58 3.15 2.76, 3.60
Type 2 diabetes mellitus
<18.5 113 1.58 1.30, 1.92 1.57 1.30, 1.91 0.99 0.82, 1.21
18.5–22.9 1,080 1.00 Referent 1.00 Referent 1.00 Referent
23.0–24.9 613 1.03 0.93, 1.14 1.03 0.93, 1.14 1.36 1.23, 1.50
25.0–27.4 603 1.67 1.51, 1.85 1.66 1.50, 1.84 2.33 2.11, 2.58
27.5–29.9 207 2.76 2.38, 3.20 2.64 2.27, 3.06 4.18 3.59, 4.87
≥30.0 119 3.53 2.92, 4.27 3.34 2.76, 4.04 5.96 4.91, 7.24
Cardiovascular disease
<18.5 182 1.20 1.03, 1.40 1.19 1.02, 1.38 1.01 0.87, 1.18
18.5–22.9 2,163 1.00 Referent 1.00 Referent 1.00 Referent
23.0–24.9 1,108 1.03 0.96, 1.11 1.04 0.96, 1.12 1.14 1.06, 1.23
25.0–27.4 697 1.11 1.02, 1.21 1.10 1.01, 1.20 1.24 1.14, 1.36
27.5–29.9 192 1.52 1.31, 1.76 1.49 1.28, 1.72 1.73 1.48, 2.02
≥30.0 99 1.77 1.45, 2.16 1.72 1.40, 2.10 2.01 1.63, 2.48
Obesity-related cancerf
<18.5 58 1.16 0.89, 1.52 1.14 0.87, 1.49 1.07 0.81, 1.40
18.5–22.9 715 1.00 Referent 1.00 Referent 1.00 Referent
23.0–24.9 421 1.14 1.01, 1.29 1.15 1.02, 1.30 1.20 1.06, 1.36
25.0–27.4 268 1.23 1.07, 1.42 1.24 1.08, 1.43 1.33 1.15, 1.54
27.5–29.9 60 1.36 1.04, 1.77 1.36 1.04, 1.77 1.53 1.17, 2.02
≥30.0 37 1.88 1.35, 2.62 1.90 1.37, 2.65 2.21 1.56, 3.12
All-cause mortality
<18.5 421 1.21 1.09, 1.33 1.18 1.07, 1.31 1.07 0.97, 1.18
18.5–22.9 4,602 1.00 Referent 1.00 Referent 1.00 Referent
23.0–24.9 2,115 1.02 0.97, 1.07 1.02 0.97, 1.08 1.07 1.01, 1.13
25.0–27.4 1,385 1.15 1.08, 1.22 1.15 1.08, 1.22 1.20 1.13, 1.29
27.5–29.9 347 1.40 1.26, 1.56 1.38 1.24, 1.54 1.44 1.28, 1.61
≥30.0 191 1.73 1.50, 2.00 1.73 1.50, 2.01 1.79 1.54, 2.09

Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio.
a
Weight (kg)/height (m)2. Trend of linearity over BMI categories was tested by modeling the median values of BMI of each category as a
continuous variable in all presented models (P < 0.001 for all models).
b
Additionally adjusted for cigarette smoking (categories of never smoker, past smoker, and current smoker of 1–4, 5–14, 15–24, 25–34, 35–44,
or ≥45 or more cigarettes/day), alcohol consumption (0, 0–5, 5–10, 10–20, 20–30, 30–45, or ≥45 g/day), and family history of myocardial infarction,
colon cancer, and diabetes.
c
Additionally adjusted for weight change between the age of 21 years and the study baseline (<−5.0, −5.0 to −2.5, −2.4 to 2.4, 2.5 to 4.9, 5.0 to
9.9, 10.0 to 14.9, and ≥15.0 kg).
d
Type 2 diabetes mellitus, cardiovascular disease, or obesity-related cancer (colorectal, renal, pancreatic, and esophageal cancers).
e
Person-time values of obesity-related diseases by category of BMI were 22,168 person-years, 346,976 person-years, 198,378 person-years,
121,347 person-years, 25,612 person-years, and 11,604 person-years, respectively. Incidence rates were 1,592 per 100,000 person-years, 1,144
per 100,000 person-years, 1,081 per 100,000 person-years, 1,295 per 100,000 person-years, 1,796 per 100,000 person-years, and 2,215 per
100,000 person-years, respectively.
f
Colorectal, renal, pancreatic, and esophageal cancers.

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 1358 de Mutsert et al.

Table 3. Hazard Ratios of Fatal and Nonfatal Obesity-Related Diseases by Weight Change Between 21 Years of Age and 1986a (n = 39,909),
Health Professionals Follow-up Study, United States, 1986–2008

Disease and No. of Age-Adjusted Multivariate HRc Adjusted for BMId

95% CI 95% CI 95% CI
Weight Change, kg Events HR HRb at Age 21 Years
e,f
Obesity-related diseases
<−5.0 313 1.18 1.04, 1.34 1.15 1.01, 1.30 0.77 0.68, 0.88
−5.0 to −2.5 207 0.96 0.83, 1.11 0.95 0.82, 1.10 0.81 0.70, 0.94
−2.4 to 2.4 1,089 1.00 Referent 1.00 Referent 1.00 Referent
2.5 to 4.9 969 1.07 0.98, 1.17 1.08 0.99, 1.17 1.12 1.03, 1.22
5.0 to 9.9 2,051 1.34 1.25, 1.44 1.34 1.24, 1.44 1.41 1.31, 1.52
10.0 to 14.9 1,644 1.58 1.46, 1.70 1.57 1.45, 1.70 1.72 1.59, 1.86
≥15.0 2,482 2.24 2.09, 2.41 2.21 2.06, 2.37 2.45 2.27, 2.63
Type 2 diabetes mellitus
<−5.0 54 1.01 0.75, 1.36 0.98 0.73, 1.32 0.50 0.37, 0.68
−5.0 to −2.5 40 0.93 0.66, 1.30 0.92 0.66, 1.29 0.69 0.49, 0.97
−2.4 to 2.4 223 1.00 Referent 1.00 Referent 1.00 Referent
2.5 to 4.9 211 1.17 0.97, 1.41 1.17 0.971.42 1.26 1.04, 1.52
5.0 to 9.9 605 2.02 1.74, 2.36 2.02 1.73, 2.35 2.21 1.89, 2.58
10.0 to 14.9 525 2.67 2.28, 3.12 2.67 2.28, 3.13 3.07 2.63, 3.60
≥15.0 1,077 5.28 4.57, 6.11 5.19 4.49, 6.00 6.02 5.20, 7.00
Cardiovascular disease
<−5.0 193 1.31 1.12, 1.54 1.28 1.08, 1.50 1.02 0.86, 1.21
−5.0 to −2.5 125 1.03 0.85, 1.25 1.02 0.84, 1.24 0.94 0.77, 1.14
−2.4 to 2.4 599 1.00 Referent 1.00 Referent 1.00 Referent
2.5 to 4.9 543 1.07 0.96, 1.21 1.08 0.96, 1.22 1.11 0.98, 1.24
5.0 to 9.9 1,059 1.23 1.11, 1.36 1.23 1.11, 1.36 1.27 1.15, 1.40
10.0 to 14.9 847 1.41 1.27, 1.57 1.41 1.27, 1.57 1.48 1.33, 1.65
≥15.0 1,075 1.68 1.51, 1.85 1.66 1.50, 1.84 1.76 1.59, 1.95
Table continues

baseline more often had a family history of disease, were
more likely to be former smokers, and were less likely to
be physically active at baseline (Table 1).
During a maximum of 22 years (726,084 person-years) of
follow-up, we documented 2,970 incident cases of type 2 di-
abetes mellitus, 4,861 of CVD, and 1,741 of obesity-related
cancers. In total, 793 men suffered from multiple events, 24
of whom experienced all 3 events. As a result, the composite
end point included 8,755 first events. In the cause-specific
analyses, we additionally censored 20 cases who reported 2
events in the same month of follow-up (i.e., it was unknown
which of the events occurred first).
Tables 2 and 3 show the incidence rates and hazard ratios
of the composite end point and the cause-specific analyses
according to BMI at 21 years of age and subsequent weight
change. As compared with having a BMI of 18.5–22.9 at 21
years of age, having a BMI of 23 or higher was consistently
associated with higher risks of all 3 chronic diseases: type 2
diabetes mellitus, CVD, and cancer (Table 2). A weight gain
of 2.5 kg or more during adulthood was associated with a
higher risk of both type 2 diabetes mellitus and CVD, but
the association between weight gain and obesity-related can-
cers was weaker (Table 3). For both BMI at age 21 years and
adult weight gain, associations for type 2 diabetes were stron-
ger than those for CVD or cancer. With regard to all-cause
mortality, having a BMI at age 21 years of 25 or higher
was associated with a higher risk than was having a BMI be-
tween 18.5 and 22.9. A weight gain of 10 kg or more was as-
sociated with a higher risk of all-cause mortality during
follow-up as compared with having a stable weight through-
out adulthood (Tables 2 and 3). Weight loss during adulthood
was associated with lower risks of type 2 diabetes mellitus
and cancer but not with lower risks of CVD and all-cause
mortality (Table 3). Results after stratification for BMI at
21 years of age are shown in Web Table 1 and described in
Web Appendix 2.
After excluding the weight change that occurred during the
5 years before baseline, adult weight gain became more
strongly associated with a higher risk of obesity-related cancer;
hazard ratios were 1.16 (95% CI: 0.98, 1.37) for a weight gain
of 10.0–14.9 kg and 1.46 (95% CI: 1.23, 1.73) for a weight
gain of 15 kg or more as compared with having a stable
weight. Associations between adult weight change and the
other disease outcomes remained similar in this sensitivity
analysis (data not shown). Restriction of the study population
to never smokers (n = 18,319), additional adjustment for

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 Overweight, Weight Change, and Chronic Disease Risk 1359

Table 3. Continued

Disease and No. of Age-Adjusted Multivariate HRc Adjusted for BMId

95% CI 95% CI 95% CI
Weight Change, kg Events HR HRb at Age 21 Years

Obesity-related cancerg
<−5.0 64 0.99 0.75, 1.30 0.97 0.74, 1.27 0.78 0.58, 1.03
−5.0 to −2.5 42 0.79 0.57, 1.10 0.80 0.58, 1.10 0.73 0.53, 1.01
−2.4 to 2.4 266 1.00 Referent 1.00 Referent 1.00 Referent
2.5 to 4.9 214 0.96 0.81, 1.16 0.96 0.80, 1.15 0.98 0.82, 1.18
5.0 to 9.9 382 1.01 0.86, 1.18 0.99 0.85, 1.16 1.03 0.88, 1.21
10.0 to 14.9 269 1.03 0.87, 1.22 1.01 0.85, 1.20 1.07 0.90, 1.27
≥15.0 322 1.15 0.97, 1.35 1.11 0.94, 1.31 1.19 1.00, 1.41
All-cause mortality
<−5.0 422 1.29 1.16, 1.44 1.27 1.14, 1.42 1.07 0.96, 1.21
−5.0 to −2.5 283 1.04 0.91, 1.18 1.04 0.91, 1.18 0.98 0.86, 1.11
−2.4 to 2.4 1,311 1.00 Referent 1.00 Referent 1.00 Referent
2.5 to 4.9 1,097 0.96 0.88, 1.04 0.96 0.89, 1.04 0.98 0.90, 1.06
5.0 to 9.9 2,095 1.02 0.95, 1.09 1.02 0.95, 1.09 1.05 0.98, 1.12
10.0 to 14.9 1,609 1.06 0.99, 1.15 1.05 0.98, 1.13 1.09 1.02, 1.18
≥15.0 2,244 1.28 1.19, 1.37 1.26 1.18, 1.35 1.31 1.22, 1.41

Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio.
a
Trend of linearity over weight-change categories was tested by modeling the median values of BMI of each category as continuous variable in
all presented models (P < 0.001 for all models, except for age-adjusted model of obesity-related cancer, for which P = 0.009).
b
Additionally adjusted for cigarette smoking (categories of never smoker, past smoker, and current smoker of 1–4, 5–14, 15–24, 25–34, 35–44,
or ≥45 or more cigarettes/day), alcohol consumption (0, 0–5, 5–10, 10–20, 20–30, 30–45, or ≥45 g/day), and family history of myocardial infarction,
colon cancer, and diabetes.
c
Additionally adjusted for BMI at 21 years of age (<18.5, 18.5–22.9, 23–24.9, 25.0–27.4, 27.5–29.9, and ≥30).
d
Weight (kg)/height (m)2.
e
Type 2 diabetes mellitus, cardiovascular disease, or obesity-related cancer (colorectal, renal, pancreatic, and esophageal cancers).
f
Person-time values of obesity-related diseases by category of weight change were 32,878 person-years, 26,736 person-years, 139,634
person-years, 111,889 person-years, 180,498 person-years, 116,565 person-years, and 117,884 person-years, respectively. Incidence rates
were 952, 774 per 100,000 person-years, 780, 866 per 100,000 person-years, 1,136 per 100,000 person-years, 1,410 per 100,000
person-years, and 2,105 per 100,000 person-years, respectively.
g
Colorectal, renal, pancreatic, and esophageal cancers.

physical activity level and dietary factors at the baseline of the
study, or exclusion of esophageal cancer from the composite
outcome did not materially alter the results (data not shown).
We evaluated the joint associations of BMI at the age of 21
years and adult weight change with the risk of chronic dis-
eases (Figure 1A). First, we compared men with different
BMIs at age 21 years who kept a stable weight throughout
adulthood. As compared with a stable BMI between 18.5
and 22.9 throughout adulthood, a stable BMI of 23–24.9
was associated with a hazard ratio of 1.47 (95% CI: 1.27,
1.71) and a stable BMI of 25 or more was associated with a
hazard ratio of 1.96 (95% CI: 1.69, 2.28). Second, we evalu-
ated whether adult weight gain was associated with a higher
risk of chronic disease even in men who had a low BMI at age
21 years. As compared with a stable low BMI (18.5–22.9)
throughout adulthood, a weight gain of 2.5–9.9 kg was asso-
ciated with a hazard ratio of 1.47 (95% CI: 1.30, 1.66), and a
weight gain of more than 10 kg with a hazard ratio of 2.35
(95% CI: 2.09, 2.64) in men with a BMI of 18.5–22.9 at
age 21 years. Third, we evaluated the combination of being
overweight at 21 years of age and having a large weight gain.
As compared with men who had a stable low BMI (18.5–
22.9), men who had a BMI of 25 or more at 21 years of
age and who gained more than 10 kg had a hazard ratio of
4.24 (95% CI: 3.73, 4.83). Overall, the relative importance
of early weight and weight change differed by disease, with
weight gain being more strongly associated with risk of type
2 diabetes mellitus and BMI in early adulthood more strongly
associated with risk of cancers (Figure 1B–D). A stable low
BMI was consistently associated with lowest risk of chronic
diseases.
Because the studied associations differed by age (P <
0.001 for interaction), we also present the results stratified
by baseline age (Figures 2–4). Although it must be noted
that the patterns of associations for BMI at age 21 years
and adult weight change in relation to risk of chronic diseases
were consistent in the different age groups, baseline age cer-
tainly influenced the magnitude of the observed associations.
With regard to relative risks, hazard ratios for chronic dis-
eases were generally greater in younger men than in older
men (Figures 3 and 4). However, the absolute incidence
rates show that the excess incidence of the chronic diseases

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 1360 de Mutsert et al.

A) B)

16 16

10 10

Hazard Ratio
Hazard Ratio

6 6
4 4
3 3
2 2
1 1
18.5–22.9 23.0–24.9 ≥25.0 18.5–22.9 23.0–24.9 ≥25.0
BMI Category BMI Category

C) D)

16 16
10 10
Hazard Ratio

Hazard Ratio
6 6
4 4
3 3
2 2
1 1
18.5–22.9 23.0–24.9 ≥25.0 18.5–22.9 23.0–24.9 ≥25.0
BMI Category BMI Category

Figure 1. Joint associations between body mass index (BMI) at 21 years of age and adult weight change with fatal and nonfatal obesity-related
diseases (A), type 2 diabetes mellitus (B), cardiovascular disease (C), and obesity-related cancer (D) among 38,613 participants in the Health Pro-
fessionals Follow-up Study, United States, 1986–2008. Obesity-related diseases included type 2 diabetes mellitus, cardiovascular disease, and
obesity-related cancer (colorectal, renal, pancreatic, and esophageal cancers). Analyses excluded participants with a BMI <18.5 (n = 1,296). Values
are hazard ratios from Cox proportional hazards model. Models were adjusted for cigarette smoking (categories of never smoker, past smoker, and
current smoker of 1–4, 5–14, 15–24, 25–34, 35–44, or ≥45 or more cigarettes/day), alcohol consumption (0, 0–5, 5–10, 10–20, 20–30, 30–45, or
≥45 g/day), and family history of myocardial infarction, colon cancer, and diabetes. , an adult weight change <−2.5 kg; , adult weight change of
−2.4 to 2.4 kg; , adult weight change of 2.5 to 9.9 kg; and , adult weight change ≥10.0 kg. Participants with a BMI of 18.5–22.9 at 21 years of age
who maintained a stable weight are the reference category. Bars, 95% confidence intervals.

related to higher BMI at 21 years of age and weight gain dur-
ing adulthood was greatest in men older than 65 years at base-
line (Figure 2).
DISCUSSION
In the present large cohort study of men from the United
States, moderate overweight in early adulthood and adult
weight gain were substantially and independently associated
with a higher risk of major chronic diseases, including type 2
diabetes mellitus, CVD (myocardial infarction and stroke),
and cancers of the colon, rectum, kidney, pancreas, and
esophagus during 22 years of follow-up. The relative impor-
tance of BMI in early adulthood and weight change appeared
to differ by disease. Cancers were more strongly associated
with BMI in early adulthood, whereas type 2 diabetes melli-
tus was more strongly associated with adult weight gain.
Moderate overweight in early adulthood and adult weight
gain were also associated with a higher risk of all-cause mor-
tality. We furthermore observed that men who were lean at
the age of 21 years and did not gain weight during adulthood
consistently had the lowest risk of the composite outcome
and the separate chronic diseases.
To evaluate the overall health impact of adiposity in early
adulthood and weight change during adulthood, we used a
composite outcome that included several major chronic dis-
eases. Most other studies investigating associations of adi-
posity in adolescence or adult weight change were limited
to 1 or a few health outcomes. Obesity in adolescence and
early adulthood has been associated with increased risks of
type 2 diabetes (13, 15), coronary heart disease (15, 17,
31–35), and stroke (32). The health consequences of moder-
ate overweight in early adulthood are less clear. A study of 3
historical cohorts in the United Kingdom reported no signifi-
cant association of being overweight early in life with future
ischemic heart disease and stroke, but that study included
very few overweight participants (14). The Harvard Growth
Study investigated a variety of diseases related to being over-
weight in adolescence and reported higher risks of coronary
heart disease and colorectal cancer in men (3). Whereas the
health effects of adult overweight have been controversial

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 Overweight, Weight Change, and Chronic Disease Risk 1361

A) B)
5,000 5,000

4,000 4,000

Incidence Rate per

Incidence Rate per

100,000 PY
100,000 PY

3,000 3,000

2,000 2,000

1,000 1,000

18.5–22.9 23.0–24.9 ≥25.0 18.5–22.9 23.0–24.9 ≥25.0

BMI Category BMI Category

C)
5,000

4,000
Incidence Rate per
100,000 PY

3,000

2,000

1,000

18.5–22.9 23.0–24.9 ≥25.0

BMI Category

Figure 2. Incidence rates per 100,000 person-years (PY) of fatal and nonfatal obesity-related diseases among 38,613 participants in the Health
Professionals Follow-up Study, United States, 1986–2008. Rates are displayed by body mass index (BMI) at 21 years of age and adult weight
change per age group: A) less than 50 years of age at baseline (n = 15,692), B) 50–64.9 years of age at baseline (n = 17,190) and C) 65 years
of age or older at baseline (n = 5,731). , an adult weight change <−2.5 kg; , adult weight change of −2.4 to 2.4 kg; , adult weight change of
2.5 to 9.9 kg; and , adult weight change ≥10.0 kg. Analyses excluded participants with a BMI less than 18.5 (n = 1,296). Obesity-related diseases
included fatal or nonfatal type 2 diabetes, cardiovascular diseases, or obesity-related cancer (colorectal, renal, pancreatic, and esophageal can-
cers). Bars, 95% confidence intervals.

(36, 37), 3 mega studies of prospective cohorts that each in-
cluded approximately a million participants consistently
showed that mortality risk was generally lowest with an
adult BMI between 20 and 25 in white adults (38, 39). Our
results indicate that BMI values in early adulthood at the
higher end of these ranges are already associated with sub-
stantially higher risks of morbidity and mortality as com-
pared with lower BMI values.
In our study, even modest adult weight gain was associated
with higher risks of CVD and type 2 diabetes mellitus, with a
particular strong association for diabetes. Large adult weight
gain (more than 10–15 kg) was also associated with a higher
risk of obesity-related cancers and all-cause mortality. These
results are in agreement with those from previous studies,
some in the same cohort, that showed that weight gain was
associated with increased risks of coronary heart disease
(11, 40–43), stroke (16, 41), and most strongly type 2 diabe-
tes mellitus (13, 19, 20). Weight gain during adulthood
has also been associated with increased risks of all-cause
mortality (5, 11, 21) and mortality due to CVD (9, 18, 21)
and certain cancers (9, 18, 21). Our results are also consistent
with those from a recent study that showed that coronary
heart disease was more closely associated with adolescent
BMI and that diabetes was mainly associated with having a
high BMI close to the time of diagnosis (16). This supports
the hypothesis that processes that cause cancer and athero-
sclerosis are gradual, whereas the effects of adiposity on in-
sulin resistance and type 2 diabetes may occur more rapidly.
We observed a lower risk of obesity-related diseases asso-
ciated with a weight loss of 2.5 kg or more during adulthood
compared with keeping a stable weight. This reduced risk
was mainly due to a markedly lower risk of type 2 diabetes.
A recent study showed that persons who were overweight or
obese as children but who became nonobese as adults had
cardiovascular risk profiles that were similar to those of per-
sons who were never obese (44). We, however, observed that
weight loss during adulthood could not completely negate the
excess risk of type 2 diabetes, CVD, or cancer in persons who
were overweight in early adulthood.
The associations of BMI in early adulthood and subsequent
weight change with risk of obesity-related diseases appeared to
be modified by age, with higher relative risks in younger par-
ticipants and higher incidence rates in participants who were
older at baseline. Strengths of the present study include the
large sample size, the prospective study design, the long
follow-up period, and the high response rate during follow-up.

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 1362 de Mutsert et al.

A) B)
10.0 10.0
6.3 6.3
4.0 4.0
2.5 2.5
Hazard Ratio

Hazard Ratio
1.6 1.6
1.0 1.0
0.6 0.6
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
.5

.9

.9

.4

.9

.5

.9

.9

.4

.9

.0
0.
8

22

24

27

29

22

24

27

29

0
<1

≥3

<1

≥3
–

–
.5

.0

.0

.5

.5

.0

.0

.5
18

23

25

27

18

23

25

27
BMI Category BMI Category

C) D)
10.0 10.0
6.3 6.3
4.0 4.0
2.5 2.5
Hazard Ratio

Hazard Ratio
1.6 1.6
1.0 1.0
0.6 0.6
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
5

0
8.

2.

4.

7.

9.

0.

8.

2.

4.

7.

9.

0.
<1

–2

–2

–2

–2

≥3

<1

–2

–2

–2

–2

≥3
.5

.0

.0

.5

.5

.0

.0

.5
18

23

25

27

18

23

25

27
BMI Category BMI Category

Figure 3. Hazard ratios of fatal and nonfatal obesity-related diseases (A), type 2 diabetes mellitus (B), cardiovascular disease (C), and
obesity-related cancer (D) by body mass index (BMI) at 21 years of age among 39,909 participants in the Health Professionals Follow-up Study,
United States, 1986–2008. ▪, <50.0 years of age at baseline (n = 16,064); ▴, 50.0–64.9 years of age at baseline (n = 17,777); ▾, ≥65.0 years of age
at baseline (n = 6,068). Obesity-related diseases included type 2 diabetes mellitus, cardiovascular diseases, and obesity-related cancer (colorectal,
renal, pancreatic, and esophageal cancers). Models were adjusted for weight change between the age of 21 years and the baseline of the study
(<−5.0, −5.0 to −2.5, −2.4 to 2.4, 2.5 to 4.9, 5.0 to 9.9, 10.0 to 14.9, and ≥15.0 kg), cigarette smoking (categories of never smoker, past smoker, and
current smoker of 1–4, 5–14, 15–24, 25–34, 35–44, or ≥45 or more cigarettes/day), alcohol consumption (0, 0–5, 5–10, 10–20, 20–30, 30–45, or
≥45 g/day), and family history of myocardial infarction, colon cancer, and diabetes. Bars, 95% confidence intervals.

Our study also has several potential limitations that need to be
considered. First, our analyses relied on self-reported and re-
called weight instead of measured weight. Validation studies
in this cohort and other cohorts, however, have shown high
accuracy of self-reported and recalled weight as compared
with measured weight (22, 45–47). The tendency to slightly
underreport body weight may have resulted in some overesti-
mation of disease risks in the higher BMI categories in our
study (48). However, because underreporting of body weight
occurs mainly in obese men (48), this may not have affected
our results in the nonobese categories. Second, we were not
able to adjust for lifestyle variables at 21 years of age. Adjust-
ment for lifestyle at baseline did not appreciably change the
results, but we cannot fully exclude the possibility of residual
confounding due to effects of lifestyle at age 21 that are
independent of lifestyle at baseline. Third, BMI is not a perfect
measure of adiposity because it also reflects variation in lean
body mass. In particular, adult weight loss until the baseline of
our study may have been affected by poor health (49). Not con-
sidering the weight change that occurred during the 5 years be-
fore baseline did not materially change the associations
between weight change and the combined obesity-related dis-
eases, whereas the association for obesity-related cancers even
became stronger. Fourth, between participants, there was a
substantial difference in the number of years between age 21
years and the baseline of the study. However, our results were
essentially the same for younger and older men. Hence, our
study provides information of the impact of adiposity on future
morbidity and mortality in men who reach middle age. Finally,
our study population included men who were predominantly

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 Overweight, Weight Change, and Chronic Disease Risk 1363

A) B)
10.0 10.0
6.3 6.3
4.0 4.0
Hazard Ratio 2.5 2.5

Hazard Ratio
1.6 1.6
1.0 1.0
0.6 0.6
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
.0

.5

.9

.5

.9
.0

0
2.

4.

9.

5.

2.

4.

9.

5.
5

–2

14

–2

14
5
<–

<–
≥1

≥1
to

to

to

to

to

to
to

to

to

to
.4

.4

0
.0

2.

5.

.0

.0

2.

5.

.0
–2

–2
–5

10

–5

10
Weight Change, kg Weight Change, kg

C) D)
10.0 10.0
6.3 6.3
4.0 4.0
2.5 2.5
Hazard Ratio

1.6 Hazard Ratio 1.6

1.0 1.0
0.6 0.6
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
.0

.5

.9

.0

.5

.9
0

0
2.

4.

9.

2.

4.

9.
5.

5.
5

–2

14

–2

14
<–

<–
≥1

≥1
to

to

to

to

to

to
to

to

to

to
.4

.4

0
.0

2.

5.

.0

.0

2.

5.

.0
–2

–2
–5

10

–5

10
Weight Change, kg Weight Change, kg

Figure 4. Hazard ratios of fatal and nonfatal obesity-related diseases (A), type 2 diabetes mellitus (B), cardiovascular disease (C), and
obesity-related cancer (D) by adult weight change in kilograms per age group among 39,909 participants in the Health Professionals Follow-up
Study, United States, 1986–2008. ▪, <50 years of age at baseline (n = 16,064); ▴, 50–65 years of age at baseline (n = 17,777); ▾, ≥65 years of
age at baseline (n = 6,068). Obesity-related diseases included type 2 diabetes mellitus, cardiovascular diseases, and obesity-related cancer (co-
lorectal, renal, pancreatic, and esophageal cancers). Models were adjusted for BMI at 21 years of age (<18.5, 18.5–22.9, 23–24.9, 25–27.4, 27.5–
29.9, ≥30), cigarette smoking (categories of never smoker, past smoker, and current smoker of 1–4, 5–14, 15–24, 25–34, 35–44, or ≥45 or more
cigarettes/day), alcohol consumption (0, 0–5, 5–10, 10–20, 20–30, 30–45, or ≥45 g/day), and family history of myocardial infarction, colon cancer,
and diabetes. Bars, 95% confidence intervals.

white, and our findings thus require confirmation in women
and in other ethnic groups.
Overweight in childhood and adolescence often continues
into adulthood (1, 50, 51) and is associated with cardiovascular
risk factors (10, 52–54). The full impact of the current in-
crease in prevalence of childhood obesity can be expected
to only become evident decades later with a high incidence
of chronic disease and reduced life expectancy (55).
Our results indicate that weight management in adulthood
should not be restricted to high-risk groups with obesity but
should also target persons who are modestly overweight. Our
results furthermore imply that weight management in adult-
hood may be more effective for reducing the risk of type 2
diabetes mellitus but that for the risk of cancer in men,
BMI in early adulthood is more important than subsequent
weight change. For the reduction of the overall risk of chronic
diseases, our results underscore the pivotal importance of
prevention of excess weight gain in both adolescence and
adulthood.
ACKNOWLEDGMENTS
Author affiliations: Department of Clinical Epidemiology,
Leiden University and Medical Center, Leiden, the Nether-
lands (Renée de Mutsert); Department of Nutrition, Harvard
School of Public Health, Boston, Massachusetts (Renée

Am J Epidemiol. 2014;179(11):1353–1365

Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075

by Ldaseptivany@gmail.com
on 02 December 2017
 1364 de Mutsert et al.
de Mutsert, Qi Sun, Walter C. Willett, Frank B. Hu, Rob
M. van Dam); Channing Division of Network Medicine, De-
partment of Medicine, Brigham and Women’s Hospital and
Harvard Medical School, Boston, Massachusetts (Qi Sun,
Walter C. Willett, Frank B. Hu); Department of Epidemiol-
ogy, Harvard School of Public Health, Boston, Massachu-
setts (Walter C. Willett, Frank B. Hu); and Saw Swee Hock
School of Public Health and Department of Medicine, Yong
Loo Lin School of Medicine, National University of Singa-
pore and National University Health System, Singapore
(Rob M. van Dam).
This work was supported by the National Institutes of
Health (grants PO1-CA055075, DK58845, P30 DK46200,
and U54CA155626), the Netherlands Organization for Sci-
entific Research (NWO) (Rubicon fellowship to Renée de
Mutsert), and the National Heart, Lung, and Blood Institute
(career development award K99HL098459 to Qi Sun).
We thank the staff of the Health Professionals Follow-Up
Study for their valuable contributions.
Preliminary results of this study were presented at the 17th
European Congress of Obesity, Amsterdam, The Nether-
lands, May 6–9, 2009, and published in abstract form
(Obes Facts. 2009;2(suppl 2):10).
Conflict of interest: none declared.
REFERENCES
1. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of high
body mass index in US children and adolescents, 2007–2008.
JAMA. 2010;303(3):242–249.
2. Hoffmans MD, Kromhout D, de Lezenne CC. The impact of
body mass index of 78,612 18-year old Dutch men on 32-year
mortality from all causes. J Clin Epidemiol. 1988;41(8):
749–756.
3. Must A, Jacques PF, Dallal GE, et al. Long-term morbidity and
mortality of overweight adolescents. A follow-up of the
Harvard Growth Study of 1922 to 1935. N Engl J Med. 1992;
327(19):1350–1355.
4. Stevens J, Truesdale KP, Wang CH, et al. Body mass index at
age 25 and all-cause mortality in whites and African
Americans: The Atherosclerosis Risk in Communities Study.
J Adolesc Health. 2012;50(3):221–227.
5. Rosengren A, Wedel H, Wilhelmsen L. Body weight and
weight gain during adult life in men in relation to coronary heart
disease and mortality. A prospective population study. Eur
Heart J. 1999;20(4):269–277.
6. van Dam RM, Willett WC, Manson JE, et al. The
relationship between overweight in adolescence and
premature death in women. Ann Intern Med. 2006;145(2):
91–97.
7. Bjorge T, Engeland A, Tverdal A, et al. Body mass index in
adolescence in relation to cause-specific mortality: a follow-up
of 230,000 Norwegian adolescents. Am J Epidemiol. 2008;
168(1):30–37.
8. Engeland A, Bjørge T, Tverdal A, et al. Obesity in adolescence
and adulthood and the risk of adult mortality. Epidemiology.
2004;15(1):79–85.
9. Shimazu T, Kuriyama S, Ohmori-Matsuda K, et al. Increase in
body mass index category since age 20 years and all-cause
mortality: a prospective cohort study (the Ohsaki Study). Int
J Obes (Lond). 2009;33(4):490–496.
Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075
by Ldaseptivany@gmail.com
on 02 December 2017
10. Neovius M, Sundstrom J, Rasmussen F. Combined effects of
overweight and smoking in late adolescence on subsequent
mortality: nationwide cohort study. BMJ. 2009;338:b496.
11. Yarnell JW, Patterson CC, Thomas HF, et al. Comparison of
weight in middle age, weight at 18 years, and weight change
between, in predicting subsequent 14 year mortality and
coronary events: Caerphilly Prospective Study. J Epidemiol
Community Health. 2000;54(5):344–348.
12. Fuemmeler BF, Pendzich MK, Tercyak KP. Weight, dietary
behavior, and physical activity in childhood and adolescence:
implications for adult cancer risk. Obes Facts. 2009;2(3):
179–186.
13. Colditz GA, Willett WC, Rotnitzky A, et al. Weight gain as
a risk factor for clinical diabetes mellitus in women. Ann Intern
Med. 1995;122(7):481–486.
14. Lawlor DA, Martin RM, Gunnell D, et al. Association of body
mass index measured in childhood, adolescence, and young
adulthood with risk of ischemic heart disease and stroke:
findings from 3 historical cohort studies. Am J Clin Nutr. 2006;
83(4):767–773.
15. Tirosh A, Shai I, Afek A, et al. Adolescent BMI trajectory and
risk of diabetes versus coronary disease. N Engl J Med. 2011;
364(14):1315–1325.
16. Rexrode KM, Hennekens CH, Willett WC, et al. A prospective
study of body mass index, weight change, and risk of stroke in
women. JAMA. 1997;277(19):1539–1545.
17. Willett WC, Manson JE, Stampfer MJ, et al. Weight,
weight change, and coronary heart disease in women. Risk
within the ‘normal’ weight range. JAMA. 1995;273(6):
461–465.
18. Jeffreys M, McCarron P, Gunnell D, et al. Body mass index in
early and mid-adulthood, and subsequent mortality: a historical
cohort study. Int J Obes Relat Metab Disord. 2003;27(11):
1391–1397.
19. Koh-Banerjee P, Wang Y, Hu FB, et al. Changes in body weight
and body fat distribution as risk factors for clinical diabetes in
US men. Am J Epidemiol. 2004;159(12):1150–1159.
20. Chan JM, Rimm EB, Colditz GA, et al. Obesity, fat distribution,
and weight gain as risk factors for clinical diabetes in men.
Diabetes Care. 1994;17(9):961–969.
21. Hu FB, Willett WC, Li T, et al. Adiposity as compared with
physical activity in predicting mortality among women. N Engl
J Med. 2004;351(26):2694–2703.
22. Rimm EB, Stampfer MJ, Colditz GA, et al. Validity of
self-reported waist and hip circumferences in men and women.
Epidemiology. 1990;1(6):466–473.
23. Chasan-Taber S, Rimm EB, Stampfer MJ, et al. Reproducibility
and validity of a self-administered physical activity
questionnaire for male health professionals. Epidemiology.
1996;7(1):81–86.
24. Rimm EB, Giovannucci EL, Stampfer MJ, et al. Reproducibility
and validity of an expanded self-administered semiquantitative
food frequency questionnaire among male health professionals.
Am J Epidemiol. 1992;135(10):1114–1126.
25. James WPT, Jackson-Leach R, Mhurchu CN, et al. Overweight
and obesity (high body mass index). In: Ezzati M, Lopez AD,
Rodgers A, et al, eds. Comparative Quantification of Health
Risks: Global and Regional Burden of Disease Attributable to
Selected Major Risk Factors. Geneva, Switzerland: World
Health Organization; 2004:497–596.
26. World Cancer Research Fund/American Institute for Cancer
Research. Food, Nutrition, Physical Activity, and the
Prevention of Cancer: a Global Perspective. Washington, DC:
American Institute for Cancer Research; 2007.
27. Stampfer MJ, Willett WC, Speizer FE, et al. Test of the National
Death Index. Am J Epidemiol. 1984;119(5):837–839.
Am J Epidemiol. 2014;179(11):1353–1365

28. 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.
29. WHO Expert Consultation. Appropriate body-mass index for
Asian populations and its implications for policy and
intervention strategies. Lancet. 2004;363(9403):157–163.
30. Koller MT, Raatz H, Steyerberg EW, et al. Competing risks and
the clinical community: irrelevance or ignorance? Stat Med.
2012;31(11-12):1089–1097.
31. Bibbins-Domingo K, Coxson P, Pletcher MJ, et al. Adolescent
overweight and future adult coronary heart disease. N Engl J
Med. 2007;357(23):2371–2379.
32. Falkstedt D, Hemmingsson T, Rasmussen F, et al. Body mass
index in late adolescence and its association with coronary heart
disease and stroke in middle age among Swedish men. Int J
Obes (Lond). 2007;31(5):777–783.
33. Owen CG, Whincup PH, Orfei L, et al. Is body mass index
before middle age related to coronary heart disease risk in later
life? Evidence from observational studies. Int J Obes (Lond).
2009;33(8):866–877.
34. Gray L, Lee IM, Sesso HD, et al. Body weight in early and
mid-adulthood in relation to subsequent coronary heart disease
mortality: 80-year follow-up in the Harvard Alumni Study.
Arch Intern Med. 2011;171(19):1768–1770.
35. Galanis DJ, Harris T, Sharp DS, et al. Relative weight, weight
change, and risk of coronary heart disease in the Honolulu
Heart Program. Am J Epidemiol. 1998;147(4):379–386.
36. Flegal KM, Graubard BI, Williamson DF, et al. Excess deaths
associated with underweight, overweight, and obesity. JAMA.
2005;293(15):1861–1867.
37. Flegal KM, Kit BK, Orpana H, et al. Association of all-cause
mortality with overweight and obesity using standard body
mass index categories: a systematic review and meta-analysis.
JAMA. 2013;309(1):71–82.
38. Prospective Studies Collaboration. Body-mass index and
cause-specific mortality in 900 000 adults: collaborative
analyses of 57 prospective studies. Lancet. 2009;373(9669):
1083–1096.
39. Berrington de Gonzalez A, Hartge P, Cerhan JR, et al.
Body-mass index and mortality among 1.46 million white
adults. N Engl J Med. 2010;363(23):2211–2219.
40. Chei CL, Iso H, Yamagishi K, et al. Body mass index and
weight change since 20 years of age and risk of coronary heart
disease among Japanese: the Japan Public Health Center-Based
Study. Int J Obes (Lond). 2008;32(1):144–151.
41. Stevens J, Erber E, Truesdale KP, et al. Long- and short-term
weight change and incident coronary heart disease and ischemic
stroke: The Atherosclerosis Risk in Communities Study. Am J
Epidemiol. 2013;178(2):239–248.
Am J Epidemiol. 2014;179(11):1353–1365
Downloaded from https://academic.oup.com/aje/article-abstract/179/11/1353/2739075
by Ldaseptivany@gmail.com
on 02 December 2017
Overweight, Weight Change, and Chronic Disease Risk 1365
42. Rimm EB, Stampfer MJ, Giovannucci E, et al. Body size and fat
distribution as predictors of coronary heart disease among
middle-aged and older US men. Am J Epidemiol. 1995;
141(12):1117–1127.
43. Colditz GA, Coakley E. Weight, weight gain, activity, and
major illnesses: the Nurses’ Health Study. Int J Sports Med.
1997;18(suppl 3):S162–S170.
44. Juonala M, Magnussen CG, Berenson GS, et al. Childhood
adiposity, adult adiposity, and cardiovascular risk factors. N
Engl J Med. 2011;365(20):1876–1885.
45. Tamakoshi K, Yatsuya H, Kondo T, et al. The accuracy of
long-term recall of past body weight in Japanese adult men. Int
J Obes Relat Metab Disord. 2003;27(2):247–252.
46. Perry GS, Byers TE, Mokdad AH, et al. The validity of
self-reports of past body weights by U.S. adults. Epidemiology.
1995;6(1):61–66.
47. Must A, Willett WC, Dietz WH. Remote recall of childhood
height, weight, and body build by elderly subjects. Am J
Epidemiol. 1993;138(1):56–64.
48. McAdams MA, Van Dam RM, Hu FB. Comparison of self-
reported and measured BMI as correlates of disease markers in
US adults. Obesity (Silver Spring). 2007;15(1):188–196.
49. Allison DB, Zannolli R, Faith MS, et al. Weight loss increases
and fat loss decreases all-cause mortality rate: results from two
independent cohort studies. Int J Obes Relat Metab Disord.
1999;23(6):603–611.
50. The NS, Suchindran C, North KE, et al. Association of
adolescent obesity with risk of severe obesity in adulthood.
JAMA. 2010;304(18):2042–2047.
51. Patton GC, Coffey C, Carlin JB, et al. Overweight and obesity
between adolescence and young adulthood: a 10-year
prospective cohort study. J Adolesc Health. 2011;48(3):
275–280.
52. Lawlor DA, Benfield L, Logue J, et al. Association between
general and central adiposity in childhood, and change in these,
with cardiovascular risk factors in adolescence: prospective
cohort study. BMJ. 2010;341:c6224.
53. Tybor DJ, Lichtenstein AH, Dallal GE, et al. Independent
effects of age-related changes in waist circumference and BMI z
scores in predicting cardiovascular disease risk factors in a
prospective cohort of adolescent females. Am J Clin Nutr. 2011;
93(2):392–401.
54. Freedman DS, Katzmarzyk PT, Dietz WH, et al. Relation of
body mass index and skinfold thicknesses to cardiovascular
disease risk factors in children: the Bogalusa Heart Study. Am J
Clin Nutr. 2009;90(1):210–216.
55. Stewart ST, Cutler DM, Rosen AB. Forecasting the effects of
obesity and smoking on U.S. life expectancy. N Engl J Med.
2009;361(23):2252–2260.