Weight increase breast cancer risk

[Guest guest]

Hi All,

Increasing weight, especially in later life, seems to increase the risk of

getting

breast cancer. This may be another benefit of " staying the course " regarding

weight

changes even later in life. That decreases of weight predicted less cancer

seems to

be of note.

See the pdf-available below.

Eng SM, Gammon MD, Terry MB, Kushi LH, Teitelbaum SL, Britton JA, Neugut AI.

Body Size Changes in Relation to Postmenopausal Breast Cancer among Women on

Long

Island, New York.

Am J Epidemiol. 2005 Jun 29; [Epub ahead of print]

PMID: 15987723

Am. J. Epidemiol. 2005 162: 229-237.

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve & db=pubmed & dopt=Abstra\

ct & list_uids=15987723 & query_hl=40

INTRODUCTION

.... weight gain in adulthood, hypothesized to increase a woman's risk of

postmenopausal breast cancer (5). This factor is tantalizing from a prevention

point

of view because it is potentially modifiable by individuals.

Body weight reflects both lean mass and fat mass, whereas weight gain in

adulthood

has been shown to represent an increase in adipose or fat tissue only (6).

Estrogens

derived from aromatization of androstenedione in peripheral fat may account for

the

increased risk of breast cancer observed among postmenopausal obese women (7).

Additionally, overweight and obesity have been linked with increased insulin

levels

and insulin resistance (8). In vitro, insulin has a mitogenic effect on breast

epithelium (9), and in vivo, insulin probably stimulates cellular proliferation

via

the mediator insulin-like growth factor 1 (10). Defects in insulin and

insulin-like

growth factors have been observed in human breast cancer cell lines, the overall

effect of which has been increased cellular proliferation (11).

In the current analysis, we made use of data from a large population-based

case-control study to characterize the relation between changes in body size and

breast cancer. Because weight exerts different effects according to menopausal

status, we restricted the analysis to postmenopausal women. Changes in weight

and

relative weight during three particular time periods in adulthood were

investigated

for their effects on postmenopausal breast cancer risk, as were patterns of body

size change throughout the lifetime. We paid particular attention to evaluation

of

the perimenopausal weight trajectory in order to address the question " How do

body

size changes later in life affect breast cancer risk? " ...

MATERIALS AND METHODS

.... between August 1, 1996, and July 31, 1997. ... Subjects' ages ranged from 20

years to 98 years. ... Covariates considered independently in logistic

regression

models included variables related to demographic factors (race, education,

marital

status, Latina ethnicity, religion), reproduction (gravidity, parity, age at

first

livebirth, breastfeeding), and menstrual cycle (age at menarche). Additionally,

we

investigated use of exogenous hormones (oral contraceptives, hormone replacement

therapy), medical history (history of biopsy-proven benign breast disease,

fertility

problems, or breast cancer in a mother, sister, or daughter), and lifestyle

factors

(alcohol consumption, dietary fat intake, total caloric intake, active cigarette

smoking, and regular participation in recreational physical activity).

.... RESULTS

Static body size variables

No relation was observed between height at age 20 years and postmenopausal

breast

cancer risk (table 1). Cases had marginally lower body sizes at age 20 years

than

controls (data not shown), but there were no significant associations between

weight

at age 20 years and postmenopausal breast cancer, as table 1 shows (for the

highest

quartile of weight at age 20 years, odds ratio (OR) = 1.04, 95 percent

confidence

interval (CI): 0.80, 1.37). Greater weight during the year preceding the

reference

date was associated with increased risk of postmenopausal breast cancer in a

dose-dependent manner (p-trend = 0.0001).

TABLE 1. Age-adjusted and multivariate-adjusted odds ratios for postmenopausal

breast cancer in relation to static body size variables, Long Island Breast

Cancer

Study Project, 1996–1997

.............................................

Body size variable----Cases Controls Age-adjusted Multivariate *

---------No.% No.% OR † 95%CI † OR 95%CI

.............................................

Height (m) at age 20 years

1.21–1.57 308 31 316 32 1.00 - - 1.00 - -

1.58–1.62 274 27 261 27 1.08 0.86, 1.37 1.04 0.81, 1.34

1.63–1.67 118 12 122 12 1.00 0.74, 1.35 0.99 0.72, 1.37

1.68–1.88 302 30 283 29 1.13 0.90, 1.41 1.12 0.88, 1.44

Unknown 4 - 8 - - - - - - -

p-trend multivariate OR - - - - - - - 0.41 - -

Weight (kg) at age 20 years

36.29–49.88 289 29 301 31 1.00 - - 1.00 - -

49.89–54.42 185 19 154 16 1.23 0.94, 1.61 1.14 0.85, 1.52

54.43–58.95 298 30 294 30 1.05 0.84, 1.32 1.02 0.79, 1.31

58.96–106.14 214 22 228 23 0.98 0.77, 1.26 1.04 0.80, 1.37

Unknown 20 13

p-trend multivariate OR - - - - - - - 0.88 - -

Weight (kg) 1 year prior to reference date

36.29–58.97 182 18 224 23 1.00 - - 1.00 - -

58.98–67.59 275 27 271 28 1.28 0.98, 1.65 1.40 1.06, 1.87

67.60–77.11 260 26 266 26 1.22 0.94, 1.59 1.38 1.04, 1.83

77.12–170.55 285 29 224 23 1.64 1.26, 2.13 1.87 1.40, 2.51

Unknown 4 5

p-trend multivariate OR 0.0001

..................................

* Adjusted for age at reference date, number of pregnancies, months of hormone

replacement therapy, history of breast cancer in a mother, sister, or daughter,

and

history of benign breast disease.

† OR, odds ratio; CI, confidence interval.

Dynamic body size variables

Body size gain in early adulthood.

There were no marked differences between case and control mean values for weight

change in early adulthood (data not shown). Greater gains in weight in early

adulthood were not associated with increased risk of breast cancer (table 2).

TABLE 2. Age-adjusted and multivariate-adjusted odds ratios for postmenopausal

breast cancer in relation to weight change, Long Island Breast Cancer Study

Project,1996-1997

......................................

Body size variable Cases Controls Age-adjusted Multivariate

----No.% No.% OR * 95%CI * OR 95%CI

............................................

Weight change (kg)from age 20 years to age 30 years y

-36.29 to -0.01 60 6 81 8 0.72 0.50,1.06 0.66 0.43,1.02

0.00 271 28 259 27 1.00 - - 1.00 - -

0.01 to 2.26 219 22 182 19 1.17 0.90,1.52 1.20 0.90,1.60

2.27 to 4.53 187 19 179 18 1.04 0.80,1.37 1.09 0.82,1.47

4.54 to 7.70 102 11 116 12 0.87 0.63,1.19 0.88 0.62,1.24

7.71 to 51.16 138 14 149 16 0.95 0.71,1.27 1.05 0.76,1.45

Unknown 29 - 24 - - - - - - -

p-trend for multivariate OR 0.80

Weight change (kg) from age 20 years to 1 year prior to reference date y

44.91 to 3.01 36 4 61 6 0.64 0.39,1.05 0.55 0.32,0.96

3.00 to 3.00 103 11 119 12 1.00 - - 1.00 - -

3.01 to 7.71 141 14 156 16 1.06 0.74,1.50 1.03 0.70,1.50

7.72 to 8.15 241 24 228 24 1.25 0.90,1.72 1.18 0.83,1.68

8.16 to 14.96 209 21 206 21 1.20 0.87,1.67 1.21 0.84,1.74

14.97 to 87.09 256 26 204 21 1.51 1.09,2.08 1.58 1.11,2.26

Unknown 33 - 33 - - - - - -

p-trend multivariate OR 0.0001

Weight change (kg) from age 50 years to 1 year prior to reference date z

68.04 to 0.01 157 17 170 18 1.00 0.74,1.37 1.19 0.85,1.67

0.00 167 18 197 21 1.00 - - 1.00 - -

0.01 to 2.71 133 14 136 15 1.11 0.81,1.53 1.19 0.84,1.69

2.72 to 4.98 124 13 146 16 0.94 0.68,1.30 0.96 0.68,1.37

4.99 to 11.33 195 20 148 16 1.45 1.07,1.97 1.58 1.14,2.23

11.34 to 62.14 171 18 125 14 1.49 1.08,2.05 1.62 1.14,2.30

Unknown 59 - 68 - - - - - - -

p-trend multivariate OR 0.003

.....................................

* OR,odds ratio;CI,con & #64257;dence interval.

y Multivariate odds ratio was adjusted for age at reference date,number of

pregnancies,months of hormone replacement therapy,history of breast cancer in a

mother,sister,or daughter,history of benign breast disease,and body mass index

at

age 20 years.

z Multivariate odds ratio was adjusted for age at reference date,number of

pregnancies,months of hormone replacement therapy,history of breast cancer in a

mother,sister,or daughter,history of benign breast disease,and body mass index

at

age 50 years.

Body size gain throughout adulthood.

Greater gains in weight throughout adulthood conferred increased risks of

postmenopausal breast cancer (table 2). Compared with women who stayed within 3

kg

of their age 20 weight, those who had gained 15 or more kg (33 pounds) between

age

20 years and the year preceding the reference date had a 58 percent greater risk

of

postmenopausal breast cancer, even when results were controlled for any residual

confounding by body mass index at the beginning of the change interval, at age

20

years (OR = 1.58, 95 percent CI: 1.11, 2.26). Accounting for a woman's baseline

body

size by examining gains as a percentage of body size at age 20 also did not

yield

materially altered risk estimates (data not shown).

Body size gain during the peri- and postmenopausal years.

Change in body size during the peri- and postmenopausal years, defined as change

occurring from age 50 to 1 year prior to the reference date, was strongly

associated

with postmenopausal breast cancer (table 2). With adjustment for body mass index

at

the beginning of the change interval, at age 50 years, women in the highest

quartile

of weight gain during this time period had a multivariate-adjusted odds ratio

for

postmenopausal breast cancer of 1.62 (95 percent CI: 1.14, 2.30) compared with

women

whose weight did not change during this time period. Controlling for residual

confounding from obesity by considering a woman's baseline body size at age 20

did

not materially change these risk estimates (data not shown).

Timing of body size gain.

Since weight change during the peri- and postmenopausal years was a component of

weight change over the lifetime, we determined whether peri- and postmenopausal

change was largely responsible for the statistically significant association

between

weight gain over the lifetime and postmenopausal breast cancer by separating

weight

change over the lifetime into its component changes (data not shown). We then

determined the time period in which weight change had most affected

postmenopausal

breast cancer risk. This analysis indicated that body size gain during the peri-

and

postmenopausal years showed the strongest association with risk of

postmenopausal

breast cancer. Women in the highest quartile of weight gain from age 50 years to

1

year prior to the reference date had an odds ratio for postmenopausal breast

cancer

of 1.64 (95 percent CI: 1.19, 2.25) as compared with women who stayed within 2

kg of

their age 50 weight; the corresponding risks for women in the highest quartile

of

weight gain from age 20 to age 30, from age 30 to age 40, and from age 40 to age

50

were 1.05, 1.05, and 1.37, respectively.

Modifiers of peri- and postmenopausal body size gain.

The odds ratio for postmenopausal breast cancer for the highest quartile of body

size gain between age 50 and 1 year prior to the reference date was

significantly

increased by 102 percent among never users of hormone replacement therapy,

whereas

it was decreased by 19 percent among women who had ever used hormone replacement

therapy (table 3). The odds ratio for estrogen receptor-positive/progesterone

receptor-positive (ER+/PR+) postmenopausal breast cancer was significantly

increased

twofold in relation to the highest quartile of body size gain between age 50 and

the

year preceding the reference date (table 4). Odds ratios for postmenopausal

breast

cancer were not found to differ by race or first-degree family history of breast

cancer (data not shown). Comparison of in-situ cases with controls and invasive

cases with controls yielded similar results (data not shown).

TABLE 3. Multivariate-adjusted* odds ratio for postmenopausal breast cancer in

relation to weight change, by use/nonuse of hormone replacement therapy, Long

Island

Breast Cancer Study Project, 1996–1997

...................................................

Weight change (kg) from age 50 years to 1 year prior to reference date

..................................................

-------Use of hormone replacement therapy Never (680 cases, 670 controls) Ever

(322

cases, 319 controls)

..................................................

---------OR 95% † CI † OR 95% CI

.................................................

–68.04 to –0.01 1.10 0.74, 1.64 0.93 0.54, 1.63

0.00 1.00 - - 1.00 - -

0.01 to 2.71 1.22 0.80, 1.87 0.96 0.57, 1.63

2.72 to 4.98 0.99 0.65, 1.50 1.03 0.58, 1.83

4.99 to 11.33 1.89 1.29, 2.79 0.79 0.45, 1.40

11.34 to 62.14 2.02 1.35, 3.02 0.81 0.43, 1.53

p for interaction multivariate OR <0.01

.................................................................

* Adjusted for age at reference date, number of pregnancies, months of hormone

replacement therapy, history of breast cancer in a mother, sister, or daughter,

history of benign breast disease, and body mass index at age 50 years.

† OR, odds ratio; CI, confidence interval.

Effects of body size loss.

Women who lost weight in early adulthood had a 34 percent decreased risk of

postmenopausal breast cancer compared with women whose body size remained the

same

during this time period (table 2). This inverse association of weight loss with

breast cancer risk was slightly more pronounced for decrease over the entire

lifetime (for weight loss from age 20 years to 1 year prior to the reference

date,

OR = 0.55, 95 percent CI: 0.32, 0.96). In contrast, a decrease in body size

during

the peri- and postmenopausal years was associated with nonsignificantly

increased

risk of postmenopausal breast cancer (for weight loss, OR = 1.19, 95 percent CI:

0.85, 1.67) in comparison with women whose body size remained the same

throughout

that time interval.

Patterns of body size cycling throughout life

Results from exploratory analyses of body size patterns throughout the lifetime

were

consistent with those from analyses of dynamic body size variables (table 5).

Subjects who were consistently at or above the control median for weight when

they

were 20, 30, and 50 years old and 1 year prior to the reference date were at

22–52

percent increased risk of postmenopausal breast cancer compared with those

remaining

below the control median. Women who switched at menopause from being below the

control median body size to above the control median had higher risks of breast

cancer than women whose body size remained low throughout adulthood (OR for

weight =

1.52), but the 95 percent confidence interval for this estimate included the

null

value. In addition, an increase in postmenopausal breast cancer risk of 111

percent

was observed among those women who exhibited a fluctuating pattern of body size

throughout adulthood, although these results were not statistically significant.

TABLE 5. Multivariate-adjusted* odds ratio for postmenopausal breast cancer in

relation to pattern of body size change, Long Island Breast Cancer Study

Project,

1996–1997

....................................................

Pattern †----Cases Controls----Odds ratio 95% confidence interval

----------------No. % No. %----

...................................................

Consistently low weight 216 28 229 30 1.00 - -

Consistently high weight 240 31 227 30 1.22 0.93, 1.61

Weight high in early adulthood and early mid-adulthood, low in peri- and

postmenopausal years 63 8 73 9 0.96 0.64, 1.45

Weight low in early adulthood and early mid-adulthood, high in peri- and

postmenopausal years 58 8 53 7 1.18 0.77, 1.81

Weight high in early adulthood, low in subsequent adulthood 48 6 58 8 0.82 0.53,

1.29

Weight low in early adulthood, high in subsequent adulthood 46 6 42 6 1.23 0.75,

2.00

Weight low in early adulthood and up to menopause, high in postmenopausal years

48 6

32 4 1.52 0.92, 2.52

Weight high in early adulthood and up to menopause, low in postmenopausal years

30 4

34 4 0.97 0.56, 1.68

Fluctuating weight 22 3 12 2 2.11 1.00, 4.44

...........................................................

* Adjusted for age at reference date, number of pregnancies, lactation status,

nulliparity, history of breast cancer in a mother, sister, or daughter, and

history

of benign breast disease.

† See Materials and Methods for explanation of patterns. " High " = greater than

or

equal to the control median; " low " = less than the control median.

DISCUSSION

Results from this large case-control study of postmenopausal women support the

hypothesis that greater gains in weight over the lifetime, particularly during

the

peri- and postmenopausal years, elevate a woman's risk of postmenopausal breast

cancer. Adjusting for residual confounding from obesity by examining weight

change

as a percentage of age 20 weight did not materially alter our findings, nor did

consideration of caloric intake or dietary fat intake. Controlling for any

confounding by body mass index at the beginning of the change interval also did

not

alter our results. Relative to women whose weight remained stable over the

lifetime,

those who had gained more than 15 kg (33 pounds) since age 20 years were at a 58

percent increased risk of breast cancer. Subjects who had gained more than 11 kg

(24

pounds) since age 50 years had 1.62-fold the breast cancer risk of those whose

weight remained the same during this time period. Weight loss over the lifetime,

defined as loss during the period from age 20 years to 1 year prior to the

reference

date, was associated with a 45 percent decrease in postmenopausal breast cancer

risk.

Our observation of a 50–60 percent increased risk of postmenopausal breast

cancer

with greater weight gain over the lifetime is consistent with most (6, 19–36),

but

not all (37–39), other investigations that have examined change in body size

throughout adulthood. The current study's finding of a 1.6-fold increased risk

of

postmenopausal breast cancer related to greater body size gain in the peri- and

postmenopausal years is consistent with the two other US case-control studies

(26,

40) that have examined gain in later adulthood. Our finding of a reduced odds

ratio

of 0.55 in relation to lifetime weight loss confirms results of an earlier

report

(25) but has not been replicated by other investigators (6, 24, 26, 28, 39).

Our finding that peri- and postmenopausal weight gain increased the risk of

postmenopausal breast cancer only among women who reported never using hormone

replacement therapy supports the hypothesis that among ever users of hormone

replacement therapy, plasma estrogen levels are elevated by exogenous hormones

even

among lean women; this may mask any effect of adiposity on breast cancer risk,

since

the increases in estrogen levels brought about by pharmacologic doses are

usually

higher than those associated with obesity alone (20, 41). Thus, only among women

who

had never used hormone replacement therapy was a clear positive association

between

body size gain and breast cancer observed. Similar findings were reported in the

Nurses' Health Study on the basis of 16 years of follow-up (20), as well as in a

large Canadian population-based case-control study (30).

Greater weight gain was associated with a twofold increase in risk of ER+/PR+

postmenopausal breast cancer. This finding suggests that body size gain later in

life may preferentially lead to ER+/PR+ tumors among postmenopausal women, which

supports the hypothesis that hormone receptor status defines biologically unique

cancers with different etiologic pathways (42). However, cell sizes were

particularly small for non-ER+/PR+ cases in comparison with ER+/PR+ cases,

making it

possible that comparisons within these groups were underpowered. These results

are

similar to those from the two studies that examined the relation of body size

change

to postmenopausal breast cancer risk according to estrogen receptor/progesterone

receptor status (29, 43).

Patterns of weight cycling throughout the lifetime were examined in an

exploratory

fashion in our study. The referent group for pattern analyses was women who

remained

consistently below the median control body size across all four time periods

examined. Women who were consistently above the control median weight were at

significantly increased risk of postmenopausal breast cancer. There was also the

suggestion that women who switched at menopause from being below the control

median

body size to above the control median were also at increased risk, although cell

sizes for this pattern were small. There have been suggestions that such

fluctuation

in body size may be indicative of a preference for foods high in dietary fat or

may

adversely affect health via its effects on metabolic rate, body composition, and

fat

distribution (44, 45). While numbers for pattern analyses were small, there was

no

indication in our study that subjects with lifetime weight fluctuations had

significantly greater dietary fat intakes than those whose weight remained more

consistent (data not shown). To date, only one study has examined the effect of

weight loss followed by weight gain on risk of postmenopausal breast cancer

(26); no

relation was reported between this definition of weight cycling and breast

cancer.

Effects of recall bias may have been minimized in our study, since hypotheses

relating breast cancer to body size gain were probably not well-known at the

time

the LIBCSP interviews were conducted. Additionally, height and weight questions

would appear normal within the setting of a health study interview. Lower body

weight has been promulgated as being socially desirable and indicative of a

healthier lifestyle, making it possible that both cases and controls

underreported

their body sizes and gains. If such nondifferential misclassification of

exposure

did occur in this study, the true effect of any gains in body size may have been

attenuated toward the null value (46) or been inflated, since the exposure

variable

was not simply dichotomous (47).

Recall of body size information from the distant past is of potential concern.

An

investigation examining the long-term recall of high school height and weight

among

elderly subjects compared recalled values with measurements obtained during

adolescence (48). Recalled and measured high school weight showed good

correlation

with one another, although there was slightly more underestimation with greater

variability of high school weight among females who were obese as adolescents

compared with those who were lean. No differences in high school weight recall

were

observed by adult body size. Another measurement-related concern in this study

is

the correlation between self-reported anthropometric measurements and those

obtained

by trained experts. A study conducted among women aged 40–81 years compared

self-reported heights and weights ascertained by mailed questionnaire with

measures

obtained by a technician (49). Self-reported measures were highly correlated

with

the direct measures, body mass index derived from self-reported measures was

highly

correlated with body mass index derived from technician measures, and the

overall

accuracy of self-reported measurements did not vary with age. These two studies

suggest that remote recall and self-measurement of anthropometric variables are

reasonably accurate even when they are used to formulate derived variables.

Selection bias may have affected the current study, since the rate of

participation

was particularly low (43.3 percent) among controls over age 65 years (12). Odds

ratios could have been overestimated, if older controls who were overweight or

had

greater body size gains in adulthood were less likely to participate in the

study,

or underestimated, if these older controls were less likely to participate

because

they were less healthy overall, making it possible that their body sizes and

anthropometric gains could have been low. Recent results from the Third National

Health and Nutrition Examination Survey indicated that the prevalence of

overweight

and obesity, defined as a body mass index greater than 25.0, in US women aged 20

years or more was 50.7 percent (50). In our study, the proportion of controls

over

age 65 years with a body mass index greater than 25.0 was 54.6 percent, which

suggests that selection bias may not have played a large role.

This study had numerous strengths, including a large sample size, which

increased

the study's power to detect small associations and allowed for subgroup

analyses.

The study's population-based design and its reasonable overall response rates

(12)

reduced the likelihood of biased sample selection and increased the findings'

generalizability to all Long Island women. Use of a comprehensive

interviewer-administered questionnaire provided detailed and well-measured

information on study variables, reducing misclassification and enhancing study

precision.

We report a positive association between postmenopausal breast cancer and weight

gain throughout adulthood, particularly during the peri- and postmenopausal

years,

independent of body size at age 20 years and at the beginning of the body size

change interval. The effects of body size gain appear to be obscured by

postmenopausal hormone use. Greater body size gains may increase the risk of

ER+/PR+

breast cancer. Weight loss over the lifetime was associated with a decreased

risk of

postmenopausal breast cancer in our study population. These results, taken

together

with current trends of increasing overweight and obesity in the United States

(50,

51) and observations from prospective studies documenting greater mortality from

cancer with increased body weight (52), underscore the fact that weight gain and

obesity are significant public health issues. The current study adds to the

anthropometry and breast cancer literature by focusing on the perimenopausal

weight

trajectory, suggesting that women can still modify their breast cancer risk

later in

life by avoiding weight gain during this period. The possibility that body size

increases may preferentially lead to ER+/PR+ breast cancers postmenopausally

deserves further study.

Al Pater, PhD; email: old542000@...

____________________________________________________

Start your day with - make it your home page

http://www./r/hs