body size and body fat distribution. Relation of ...

Relation of endometrial cancer risk to past and contemporary body size and body fat distribution. C A Swanson, N Potischman, G D Wilbanks, et al. Cancer Epidemiol Biomarkers Prev 1993;2:321-327. Published online July 1, 1993.

Updated Version

Citing Articles

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cebp.aacrjournals.org/content/2/4/321

This article has been cited by 11 HighWire-hosted articles. Access the articles at: http://cebp.aacrjournals.org/content/2/4/321#related-urls

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected].

Downloaded from cebp.aacrjournals.org on July 14, 2011 Copyright © 1993 American Association for Cancer Research

Vol.

2, 321-327,

July/August

Relation

Cancer Epidemiology, Biomarkers & Prevention

1993

of Endometrial Cancer Risk to Past and Contemporary Body Size and Body Fat Distribution

Christine A. Swanson,’ Nancy Potischman, George D. Wilbanks, Leo B. Twiggs, Rodrigue Mortel, Michael L Berman, Rolland J. Barrett, Richard N. Baumgartner, and Louise A. Brinton Environmental Epidemiology Branch, National Cancer Institute, Bethesda, Maryland 20892 [C. A. S., N. P., L. A. B.]; Department of Obstetrics and Gynecology, Rush Medical College, Chicago, Illinois 60612 [G. D. W.]; Department of Obstetrics and Gynecology, University [L. B. T.];

of Minnesota Department

Medical School, of Obstetrics and

Minneapolis, Gynecology,

Minnesota 55455 Milton S. Hershey

Medical Center, Hershey, Pennsylvania 1 7033 [R. M.J; Department of Obstetrics and Gynecology, University of California at Irvine Medical Center, Irvine, California 92668 [M. L. B.]; Department of Obstetrics and Gynecology, The Bowman Gray School of Medicine, WinstonSalem, North Carolina 27103 [R. J. B.]; and Clinical Nutrition Research Laboratory, School of Medicine, University of New Mexico, Albuquerque,

New

Mexico

87131

ER. N. B.]

Abstrad

In a multicenter case-control study that included 403 cases and 297 controls, we examined the relation of past and contemporary body size, including body fat distribution, to the risk of endometrial cancer. The relative contributions of past and contemporary body size were assessed by examining weight and height histories provided by the subjeds. Anthropometric indicators thought to refled early environmental influences (e.g., height and sitting height), current weight, and fat distribution patterns were measured diredly. Height was not a risk fador for endometrial cancer, but inexplicably, sitting height was inversely associated with risk. Weight during early adulthood appeared to be diredly related to disease risk, but the association was explained by contemporary weight and thus weight gain during adulthood. While contemporary weight was associated with risk of endometrial cancer, the effed was restrided to those in the top quartile. Women whose measured weight at interview exceeded 78 kg had 2.3 times the risk of those weighing less than 58 kg (95% confidence interval, 1.4 to 3.7). Upper-body obesity (waist-to-thigh circumference ratio) was a risk fador independent of body weight. After adjustment for weight, the relative risks of endometrial cancer across increasing quartiles of upper-body obesity were 1 .0, 1 .5, 1 .8, and 2.6 (P for trend < 0.001). These data indicate that both obesity and the distribution of adipose tissue accumulated during adult life increase endometrial cancer risk substantially.

Received 1

To whom

7/8/92.

requests

for reprints

should

be addressed.

Introdudion Many epidemiological studies show that the risk of endometnial cancer increases considerably with increasing adult weight (1-4). However, in only a few studies has the effect of obesity at a young age been assessed. Blitzer et a!. (5) reported that teenage obesity predisposed to endometrial cancer, particularly when overweight persisted throughout adult life. Levi et a!. (6) made a similar observation but concluded, as have others (4, 7, 8), that body mass in the few years before diagnosis was the main determinant of risk. The role offat distribution patterns has only recently been examined in epidemiological studies of endometrial cancer. Although some investigations show little or no effect of fat distribution independent of obesity (4, 9), other studies (10, 11) indicate that the risk of the disease is markedly increased among women who tend to accumulate fat on their upper body. Height has been given less attention than weight as a possible risk factor for cancer of the endometnium, yet one early study (1 2) and a more recent report (1 3) provide evidence of a direct association. Height is of particular interest because, in some populations, adult stature may reflect early environmental influences such as nutrition. Sitting height may be particularly sensitive to early diet, and relative sitting height (i.e., the sitting height-to-standing height ratio) has been proposed as an indirect index of preadolescent nutrition. Briefly, long-waisted women tend to experience early menarche, and early maturation may reflect increased nutrition during childhood (14). A case-control study provided an opportunity to explore in detail the relation between body size and the risk of endometrial cancer. Indicators of current weight and fat distribution patterns were measured directly as were anthropometric dimensions reported to reflect early environmental influences. The relative contributions of early and contemporary body size were assessed by examining weight and height histories provided by subjects.

Methods This article is based on data obtained from women who participated in a multicenter case-control study conducted during 1987 to 1990 (1 5). The cases were women aged 20 to 74 years from five geographic areas of the United States who were newly diagnosed with endometrial cancer. A total of 498 cases was considered eligible for study. For each eligible case, we attempted to select one control matched for age (same 5-year group), race, and location of residence (based on either telephone exchange or zip code). For cases under the


321

322

Body

Size and Endometrial

Cancer

age of 65 years, controls were selected using RDD2 techniques (16). Ofthe residential households identified through RDD, an enumeration of female members was obtained for 86%. Older controls were identified by randomly selecting women from current HCFA computer tapes. A total of 1 25 of the initially selected RDD controls and 88 of the HCFA controls were eliminated because they reported hysterectomies. These women were replaced with other eligible subjects. A total of 304 RDD and 1 73 HCFA controls was enlisted. Interviews were completed with 434 of 498 eligible cases (87.1%) and with 313 of 477 eligible controls (65.6%). The primary reason for nonresponse was refusal (4.8% of cases versus 21 .8% of controls). Eligible subjects who could not be interviewed were not replaced. Two cases provided no anthropometry information, either measured or reported, and were excluded from the analysis. All cases were pathologically confirmed, with 93% of the interviewed cases having a classification of epithehal cancer. Because of distinct epidemiological charactenistics of sarcomas (1 7), this analysis focused on interviews with the 403 epithelial cancer cases and their 297 matched controls. Respondents were interviewed by uniformly trained interviewers who administered a standardized questionname which included detailed information on demographic factors, pregnancies, menstruation and menopause, contraceptive behavior, use of exogenous hormones, diet, alcohol intake, smoking, medical conditions, and family history of cancer. Included in the questionname were inquiries about perceptions of early weight and height and also questions about weight changes during adult life. For example, subjects were asked to place themselves along a continuum of weight (thin, moderately thin, average, moderately heavy, or heavy) during a preadolescent period (ages 8 to 9 years) and at 1 8 years of age. A similar set of questions was asked with respect to height. Respondents were asked to report their adult height (in feet and inches), current weight (in pounds), and usual weight for each 10-year interval starting with the third decade. To assess repeated weight gain and loss, the women were asked how many times over their lifetime they had lost as much as 20 pounds and then gained it back. Interviewers

in

each

of

the

five

centers

received

standardized training in anthropometnic techniques from one of the investigators (R. N. B.). Twice during the course of the study, the anthropometry instructor evaluated the quality of the data collected at each center. The following anthropometnic indicators were measured: standing and sitting height to the nearest 0.1 cm with a pocket stadiometer (Holtain; Senitex, Carlstadt, NJ); weight to the nearest 0.1 kg with a portable digital scale (lntegra 815; Seca, Columbia, MD); midarm, waist, and thigh

circumferences

to the

nearest

0.1 cm with

a spring-

loaded fiberglass tape (McMaster Carr Supply Company, Dayton, NJ); and subscapular, triceps, and suprapatellar skinfolds to the nearest 0.1 mm with skinfold calipers (Tanner 610, Senitex, Carlstadt, NJ). Midarm circumference was measured at the midpoint of the arm between

2

The abbreviations

Care Financing

used

are: RDD,

Administration;

random

CI, confidence

digit

dialing;

interval;

HCFA, RR, relative

Health risk

the acromion and olecranon process; waist circumference was measured at the level of the umbilicus; and thigh circumference was measured at one-third of the distance from the proximal border of the patella to the anterior iliac spine. Triceps skinfolds were measured on the posterior midline of the arm at the same level as the midarm circumference. The subscapular skinfold site was identified by locating the area just below the inferior angle of the scapula. The suprapatellar skinfold site was located about 2 cm above the proximal edge of the patella. Circumferences occasionally were measured over clothing, usually a light robe or a loose-fitting garment. Limb circumferences and skinfolds were measured on the right side of the body. Measurements were taken twice and a third measurement was taken if the difference of the first two measurements exceeded a predetermined limit. Obvious outliers were removed before calculating the mean of all replicates. Ninety-two % of the cases were measured within 3 months of diagnosis. Forty-nine % of all cases were measured before surgery. Of those measured postoperatively, most (87%) were measured within 3 months. Several indices of adiposity and adipose tissue distnibution were examined. To assess obesity, we evaluated a body mass index for women (kg/mt5) recommended by the National Center for Health Statistics (18) and the more commonly reported Quetelet’s index (kg/m2). Both indicators of weight adjusted for height gave similar resuIts

and

only

Quetelet’s

index

is reported.

Skinfold

summations (e.g., subscapular, triceps, suprapatellar) were also examined but did not provide additional information about the effects of excess weight or obesity. To assess fat distribution patterns, two indices were derived. The waist-to-thigh circumference ratio was used as an index of upper-body or android obesity (19). The subscapular-to-triceps skinfold ratio provided an index of central obesity (20), a measure of the s.c. fat on the trunk of the body versus the periphery. To estimate the risk of endometnial cancer associated with various anthropometnic variables, we calculated odds ratios as approximations of relative risks. Unconditional logistic regression was used to adjust for potential confounding variables, deriving maximum likelihood estimates of relative risks and 95% confidence intervals (21). Logistic regression was also used to test the statistical significance of interaction terms. Tests for trend in the logistic analyses were obtained by categorizing the exposure variable and treating the scored variables as continuous, after eliminating unknown values. The regression analyses included age as a continuous variable. Results were similar when age was modelled as a categorical variable (65 years). Education, smoking status, number of births, and menopausal estrogen use were entered into regression analyses as categorical variables. Further adjustment for other potential confounders (i.e., age at menarche; oral contraceptive use; age at menopause; diabetes) did not matenially alter the results.

Results The mean age of cases at interview was 59.2 years, compared with 58 years among the controls. Seventyfour % of cases were menopausal compared with 70% of controls. Cases and controls were comparable on race,


Cancer Epidemiology, Biomarkers & Prevention

Table

1 Distribution of endometrial according to other endometnial

cancer cancer

Ca ses Risk

Education 78.3

>78.3

RR’

54 64 53

55 for

Because

kg

Controls

30 62 57

participants.

for reported contemporary weight, weight in the twenties was not related to risk of endometnial cancer (data not shown). We also examined weight change from the twenties to interview (Table 6). Whether expressed as percentage of change or as an absolute increase, weight gain during adulthood was directly associated with endometrial cancer risk. While risk was elevated among women who lost weight, the confidence intervals included one. The effect of weight gain by strata of Quetelet’s index at

Measured Waist-to-thigh circumfer-

study

reliability of reporting of weight among older subjects we evaluated the associations by age strata (