Int. J. Cancer: 50,567-571 (1992) 0 1992 Wiley-Liss, Inc.
Publication of the International Union Against Cancer Publication de I’Unlon Internationale Contre le Cancer
BODY MASS AT DIFFERENT AGES AND SUBSEQUENT ENDOMETRIAL CANCER RISK Fabio LEVI’’2’5, Carlo LA VECCHIA’”, Eva NEGRI3,Fabio PARAZZINI’ and Silvia FRANCESCH14 ’Registre Vaudois des Tumeurs, C H W Falaises 1, 1005 Lausanne; 21nstituteof Social and Preventive Medicine, University of Lausanne, 1005 Lausanne, Switzerland; 31stitutodi Ricerche Farmacologiche “Mario Negti’: Via Eritrea 62, 20157 Milano; and 4Serviziodi Epidemiologia, Centro Riferimento Oncologico, 33081 Aviuno (PN), Italy. The relationship between body mass index (BMI) at different ages and subsequent endometrial-cancer risk was investigated in a multicentre case-control study conducted between 1988 and 1991 in Vaud, Switzerland, and Northern Italy on 272 histologically confirmed incident cases of endometrial cancer and 57 I controls admitted to hospital for acute, non-neoplastic conditions, unrelated to known or potential risk factors for endometrial cancer. The risk of endometrial cancer increased with increasing BMI in the 3rd decade of age (20 to 29 years), in the 5th decade (40 to 49 years) and in the 7th decade (60 to 69 years), although the risk estimates tended to be substantially higher at older ages: compared with women whose BMI (kg m-3 was < 20, the relative risks (RR) were I .8 for BMI 2 25 at age 20 to 29,2.7 for BMI 2 30 at age 40 to 49 and 3.8 at age 60 to 69. All the trends in risk were significant, except that for BMI at age 25 after allowance for current BMI. When data were examined in separate strata of current BMI, among women of normal body mass at diagnosis no significant effect of past Overweight was observed. In contrast, among subjects overweight at diagnosis, there were significant direct relationships with BMI at ages 20 to 29 and 40 to 49. To reduce endometrial cancer risk, it is therefore important to avoid obesity in later middle and older age, and the benefit can be even greater for women who were overweight at younger age.
Overweight and obesity are well-recognized risk factors for endometrial cancer. The association was consistently observed in at least 18 studies, and present both in pre- and in post-menopausal women, although the underlying biological mechanisms are likely to differ at different ages (Elwood et al., 1977; Lew and Garfinkel, 1979; La Vecchia, 1989; Parazzini et al., 1991). In fact, increased aromatization of androgens to oestrogens and reduced sex-hormone-binding globulin are the relevant factors in post-menopause, while the influence of obesity on frequency of anovulation is important in premenopause, with the consequent amenorrhoea and lutealphase progesterone deficiency (La Vecchia et al., 1984; Key and Pike, 1988; La Vecchia, 1989). It is not known whether nutritional and anthropometric factors early in life are related to the subsequent risk of endometrial cancer. Two case-control studies (Wynder et al., 1966; Koumantaki et al., 1989) and one cohort study (Tretli and Magnus, 1990) found a direct association between height (which is an indirect indicator of nutritional status in childhood and adolescence) and endometrial-cancer risk, but no relationship was observed in another case-control investigation (La Vecchia et al., 1991). Similarly inconsistent are the data on the possible relationship with weight at early age or weight changes. A cohort study of over 55,000 women among whom 133 cases of endometrial carcinoma were diagnosed showed a direct relationship with maximal weight at age 10 to 20, independent of current body weight (Blitzer et aZ., 1976). However, another historical cohort study of Hawaiian women found no clear association with weight, height or body mass at ages 10 to 29 (Le Marchand et ul., 1991), and a case-control study of endometrial cancer in women aged 45 or less found no association with weight at age 18 (Henderson et al., 1983). Likewise, a study of 200 cases and 900 controls conducted in New York in 1959-63 (Wynder et al., 1966) found that for each
weight at age 25 to 29 the risk increased with increased weight at age 50 to 59, but the reverse was not true. Nevertheless, women who had always been overweight were at greater risk than those who were overweight at diagnosis of endometrial cancer, but not at a younger age. The question has important theoretical implications for the understanding of the role of overweight, and its biological correlates, in the process of endometrial carcinogenesis, ie., whether it has essentially a late-stage effect (such as that of exogenous oestrogens) (Davis et al., 1981; La Vecchia et al., 1982, 1984; Pike, 1987; Key and Pike, 1988) or an early-stage effect also. Further, a better understanding of this time-risk relationship has practical relevance for prevention. To provide further information on these issues, we report on endometrial-cancer risk in relation to measures of body mass index at different ages, using data from a multicentre casecontrol study conducted in Switzerland and Italy. SUBJECTS AND METHODS
Since January 1988, we have been conducting a cooperative case-control study of endometrial cancer in the Swiss Canton of Vaud and in metropolitan Milan and Pordenone, FriuliVenezia Giulia Region, Northern Italy. The general design of this investigation has been already described (Levi et al., 1991). Briefly, in Vaud the recruitment of cases was populationbased, since identified cases were matched with incident cases reported to the local Cancer Registry (Levi, 1987). Overall, over 80% of identified cases were interviewed. In Milan and Pordenone, case recruitment was hospital-based, since the areas are not covered by cancer registration schemes. Thus, we had no precise information of the proportion of incident cases identified and interviewed, although the hospitals included in the study scheme included the major teaching and general hospitals in the study areas. Overall, 272 histologically confirmed endometrial cancers were interviewed. The age range was 31 to 75 (median age, 61 years). The controls were women aged 75 or less, admitted to the same networks of hospitals where cases had been identified, with primary diagnosis unrelated to any of the known or suspected risk factors for endometrial cancer or to any long-term modification in diet. Specific exclusion was made of women admitted for gynaecological, hormonal, metabolic or neoplastic diagnoses, or who had undergone hysterectomy. A total of 571 women, aged 30 to 75 (median age 59) were interviewed. Of them, 32% were admitted for traumas, 16% had non-traumatic orthopaedic diseases, 26% surgical conditions and 26% other miscellaneous disorders, including acute medical conditions, eye, nose and throat affections, dental diseases, etc. All interviews were conducted in hospital. Less than 10% of subjects approached for interview refused. The same struc5Towhom correspondence and reprint requests should be sent. Received: September 2, 1991 and in revised form October 18,1991.
568
LEV1 ET AL.
TABLE I -DISTRIBUTION OF 212 CASES OF ENDOMETRIAL CANCER AND 571 CONTROLS ACCORDING TO STUDY CENTRE, AGE, MENOPAUSAL STATUS AND CURRENT BODY MASS INDEX Endometrial cancer
Study centre Vaud Milan Pordenone Age group (years) < 45 45-54 55-64 65-74 Menopausal status Pre- in menopause Post-menopause Current body mass index (kg m-’) < 20 20-24 25-29 2 30
Controls
Number
%
Number
%
138 111 23
50.7 40.8 8.5
406 121 44
71.1 21.2 7.7
16 50 111 95
5.9 18.4 40.8 34.9
65 128 187 191
11.4 22.4 32.7 33.5
55 217
20.2 79.8
156 415
27.3 72.7
17 110 93 52
6.3 40.4 34.2 19.1
60 252 190 69
10.5 44.1 33.3 12.1
TABLE I1 - DISTRIBUTION OF 212 CASES OF ENDOMETRIAL CANCER AND 571 CONTROLS ACCORDING TO BODY MASS INDEX (BMI)I N SUBSEQUENT AGE GROUPS’
Aee 3rd decade (20 to 29) Cases
Body mass index (kg m-’) < 20 20-24 25-29 2 30 (Correlation coefficient with current BMI)
Controls
74 178 148 326 33 58 11 7 (0.45)
5th decade (40 to 49) Cases
Controls
26 61 117 296 84 146 31 34 (0.79)
7th decade (60 to 69)
Cases
6 65 46 31
Controls
25 110 100 39 (0.97)
’The sum of strata does not add up to the total because of a few missingvalues and subjects below age 20,40 or 60
(ie., BMI = 25-29) and 2.5 (95% CI 1.3 to 5.1) for severely obese ones (BMI = 2 30). The distribution of cases and controls according to body mass index in the 3rd decade (20to 29 years), the 5th (40to 49 years) and the 7th (60 to 69 years) decades is presented in Table 11, together with the correlation coefficients with current body mass index. The correlation was moderate with BMI at age 20 to 29 (r = 0.45), but rose to 0.97for BMI at age 60 to 69. The corresponding relative risks are given in Table 111. The Data analysis risk increased with increasing body mass at each separate age, Relative risk (RR) and the corresponding 95% confidence although the risk estimates tended to be systematically higher interval (CI) of endometrial cancer according to measures of at older ages: compared with women whose BMI was below 20, body mass index (BMI) (Quetelet’s index, kg m-’) were the RR was 1.8for BMI 2 25 at age 20 to 29 (data were too computed, using unconditional multiple logistic regression, scanty to assess women with BMI 2 30 in this age group), and with maximum likelihood fitting (Baker and Nelder, 1978; 2.7 and 3.8, respectively for BMI 2 30 at age 40 to 49 and at Breslow and Day, 1980). Included in the regression equations age 60 to 69.All the trends in risk were significant, except that were terms for study centre, age (in quinquennia), education for BMI at age 20 to 29 after allowance for current BMI, and at (
Publication of the International Union Against Cancer Publication de I’Unlon Internationale Contre le Cancer
BODY MASS AT DIFFERENT AGES AND SUBSEQUENT ENDOMETRIAL CANCER RISK Fabio LEVI’’2’5, Carlo LA VECCHIA’”, Eva NEGRI3,Fabio PARAZZINI’ and Silvia FRANCESCH14 ’Registre Vaudois des Tumeurs, C H W Falaises 1, 1005 Lausanne; 21nstituteof Social and Preventive Medicine, University of Lausanne, 1005 Lausanne, Switzerland; 31stitutodi Ricerche Farmacologiche “Mario Negti’: Via Eritrea 62, 20157 Milano; and 4Serviziodi Epidemiologia, Centro Riferimento Oncologico, 33081 Aviuno (PN), Italy. The relationship between body mass index (BMI) at different ages and subsequent endometrial-cancer risk was investigated in a multicentre case-control study conducted between 1988 and 1991 in Vaud, Switzerland, and Northern Italy on 272 histologically confirmed incident cases of endometrial cancer and 57 I controls admitted to hospital for acute, non-neoplastic conditions, unrelated to known or potential risk factors for endometrial cancer. The risk of endometrial cancer increased with increasing BMI in the 3rd decade of age (20 to 29 years), in the 5th decade (40 to 49 years) and in the 7th decade (60 to 69 years), although the risk estimates tended to be substantially higher at older ages: compared with women whose BMI (kg m-3 was < 20, the relative risks (RR) were I .8 for BMI 2 25 at age 20 to 29,2.7 for BMI 2 30 at age 40 to 49 and 3.8 at age 60 to 69. All the trends in risk were significant, except that for BMI at age 25 after allowance for current BMI. When data were examined in separate strata of current BMI, among women of normal body mass at diagnosis no significant effect of past Overweight was observed. In contrast, among subjects overweight at diagnosis, there were significant direct relationships with BMI at ages 20 to 29 and 40 to 49. To reduce endometrial cancer risk, it is therefore important to avoid obesity in later middle and older age, and the benefit can be even greater for women who were overweight at younger age.
Overweight and obesity are well-recognized risk factors for endometrial cancer. The association was consistently observed in at least 18 studies, and present both in pre- and in post-menopausal women, although the underlying biological mechanisms are likely to differ at different ages (Elwood et al., 1977; Lew and Garfinkel, 1979; La Vecchia, 1989; Parazzini et al., 1991). In fact, increased aromatization of androgens to oestrogens and reduced sex-hormone-binding globulin are the relevant factors in post-menopause, while the influence of obesity on frequency of anovulation is important in premenopause, with the consequent amenorrhoea and lutealphase progesterone deficiency (La Vecchia et al., 1984; Key and Pike, 1988; La Vecchia, 1989). It is not known whether nutritional and anthropometric factors early in life are related to the subsequent risk of endometrial cancer. Two case-control studies (Wynder et al., 1966; Koumantaki et al., 1989) and one cohort study (Tretli and Magnus, 1990) found a direct association between height (which is an indirect indicator of nutritional status in childhood and adolescence) and endometrial-cancer risk, but no relationship was observed in another case-control investigation (La Vecchia et al., 1991). Similarly inconsistent are the data on the possible relationship with weight at early age or weight changes. A cohort study of over 55,000 women among whom 133 cases of endometrial carcinoma were diagnosed showed a direct relationship with maximal weight at age 10 to 20, independent of current body weight (Blitzer et aZ., 1976). However, another historical cohort study of Hawaiian women found no clear association with weight, height or body mass at ages 10 to 29 (Le Marchand et ul., 1991), and a case-control study of endometrial cancer in women aged 45 or less found no association with weight at age 18 (Henderson et al., 1983). Likewise, a study of 200 cases and 900 controls conducted in New York in 1959-63 (Wynder et al., 1966) found that for each
weight at age 25 to 29 the risk increased with increased weight at age 50 to 59, but the reverse was not true. Nevertheless, women who had always been overweight were at greater risk than those who were overweight at diagnosis of endometrial cancer, but not at a younger age. The question has important theoretical implications for the understanding of the role of overweight, and its biological correlates, in the process of endometrial carcinogenesis, ie., whether it has essentially a late-stage effect (such as that of exogenous oestrogens) (Davis et al., 1981; La Vecchia et al., 1982, 1984; Pike, 1987; Key and Pike, 1988) or an early-stage effect also. Further, a better understanding of this time-risk relationship has practical relevance for prevention. To provide further information on these issues, we report on endometrial-cancer risk in relation to measures of body mass index at different ages, using data from a multicentre casecontrol study conducted in Switzerland and Italy. SUBJECTS AND METHODS
Since January 1988, we have been conducting a cooperative case-control study of endometrial cancer in the Swiss Canton of Vaud and in metropolitan Milan and Pordenone, FriuliVenezia Giulia Region, Northern Italy. The general design of this investigation has been already described (Levi et al., 1991). Briefly, in Vaud the recruitment of cases was populationbased, since identified cases were matched with incident cases reported to the local Cancer Registry (Levi, 1987). Overall, over 80% of identified cases were interviewed. In Milan and Pordenone, case recruitment was hospital-based, since the areas are not covered by cancer registration schemes. Thus, we had no precise information of the proportion of incident cases identified and interviewed, although the hospitals included in the study scheme included the major teaching and general hospitals in the study areas. Overall, 272 histologically confirmed endometrial cancers were interviewed. The age range was 31 to 75 (median age, 61 years). The controls were women aged 75 or less, admitted to the same networks of hospitals where cases had been identified, with primary diagnosis unrelated to any of the known or suspected risk factors for endometrial cancer or to any long-term modification in diet. Specific exclusion was made of women admitted for gynaecological, hormonal, metabolic or neoplastic diagnoses, or who had undergone hysterectomy. A total of 571 women, aged 30 to 75 (median age 59) were interviewed. Of them, 32% were admitted for traumas, 16% had non-traumatic orthopaedic diseases, 26% surgical conditions and 26% other miscellaneous disorders, including acute medical conditions, eye, nose and throat affections, dental diseases, etc. All interviews were conducted in hospital. Less than 10% of subjects approached for interview refused. The same struc5Towhom correspondence and reprint requests should be sent. Received: September 2, 1991 and in revised form October 18,1991.
568
LEV1 ET AL.
TABLE I -DISTRIBUTION OF 212 CASES OF ENDOMETRIAL CANCER AND 571 CONTROLS ACCORDING TO STUDY CENTRE, AGE, MENOPAUSAL STATUS AND CURRENT BODY MASS INDEX Endometrial cancer
Study centre Vaud Milan Pordenone Age group (years) < 45 45-54 55-64 65-74 Menopausal status Pre- in menopause Post-menopause Current body mass index (kg m-’) < 20 20-24 25-29 2 30
Controls
Number
%
Number
%
138 111 23
50.7 40.8 8.5
406 121 44
71.1 21.2 7.7
16 50 111 95
5.9 18.4 40.8 34.9
65 128 187 191
11.4 22.4 32.7 33.5
55 217
20.2 79.8
156 415
27.3 72.7
17 110 93 52
6.3 40.4 34.2 19.1
60 252 190 69
10.5 44.1 33.3 12.1
TABLE I1 - DISTRIBUTION OF 212 CASES OF ENDOMETRIAL CANCER AND 571 CONTROLS ACCORDING TO BODY MASS INDEX (BMI)I N SUBSEQUENT AGE GROUPS’
Aee 3rd decade (20 to 29) Cases
Body mass index (kg m-’) < 20 20-24 25-29 2 30 (Correlation coefficient with current BMI)
Controls
74 178 148 326 33 58 11 7 (0.45)
5th decade (40 to 49) Cases
Controls
26 61 117 296 84 146 31 34 (0.79)
7th decade (60 to 69)
Cases
6 65 46 31
Controls
25 110 100 39 (0.97)
’The sum of strata does not add up to the total because of a few missingvalues and subjects below age 20,40 or 60
(ie., BMI = 25-29) and 2.5 (95% CI 1.3 to 5.1) for severely obese ones (BMI = 2 30). The distribution of cases and controls according to body mass index in the 3rd decade (20to 29 years), the 5th (40to 49 years) and the 7th (60 to 69 years) decades is presented in Table 11, together with the correlation coefficients with current body mass index. The correlation was moderate with BMI at age 20 to 29 (r = 0.45), but rose to 0.97for BMI at age 60 to 69. The corresponding relative risks are given in Table 111. The Data analysis risk increased with increasing body mass at each separate age, Relative risk (RR) and the corresponding 95% confidence although the risk estimates tended to be systematically higher interval (CI) of endometrial cancer according to measures of at older ages: compared with women whose BMI was below 20, body mass index (BMI) (Quetelet’s index, kg m-’) were the RR was 1.8for BMI 2 25 at age 20 to 29 (data were too computed, using unconditional multiple logistic regression, scanty to assess women with BMI 2 30 in this age group), and with maximum likelihood fitting (Baker and Nelder, 1978; 2.7 and 3.8, respectively for BMI 2 30 at age 40 to 49 and at Breslow and Day, 1980). Included in the regression equations age 60 to 69.All the trends in risk were significant, except that were terms for study centre, age (in quinquennia), education for BMI at age 20 to 29 after allowance for current BMI, and at (
